DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of Claims
In response filed on 03 August 2026 the following changes have been made: amendments to claims 1, 5, 7, 18, and 20. Claims 3-4 have been cancelled. Claims 22-23 have been added.
Claims 1-2, 5-10, and 12-23 are currently pending and have been examined.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP §§ 706.02(l)(1) - 706.02(l)(3) for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/process/file/efs/guidance/eTD-info-I.jsp.
Claims 1-2, 5-10, and 12-23 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1-23 of App# 18/994,127 in view of Solanki et al. (US20150110368A1), Burlina et al. (US20150265144A1), and further in view of Nozaki et al. (US11024031B1). Although the claims at issue are not identical, they are not patentably distinct from each other because recite substantially similar limitations.
This is a provisional nonstatutory double patenting rejection.
The table/chart below exhibits the similarity* between the independent claims where claims 1 and 20 of the current application is a narrower variation of the claims of the reference patent.
* Similarities highlighted in BOLD
App# 18/811,687
App# 18/994,127
Claim 1:
performing a pairwise comparison of the fundus image against each example fundus image in a reference image set using one or more convolution
neural networks to determine a probability or a decision as to whether the fundus image is more
or less severe than each example fundus image in the reference image set, and generating a comparison output representing the probability or the decision for each pairwise comparison; and
flagging the fundus image as requiring referral for investigation for the disease based on results of the pairwise comparisons by mapping the generated comparison outputs to an ordered severity scale of the reference image set,
wherein the reference image set includes a plurality of example fundus images, the plurality of example fundus images ranked according to their degree of severity of disease
Claim 1:
performing a pairwise comparison of the medical data sample against each example of medical data in a reference data set using one or more machine learning algorithms to determine a difference in severity of the medical data compared to each example medical data in the reference data set; and
flagging the medical data sample as requiring referral for investigation for the disease based on results of the pairwise comparisons, wherein the reference data set includes a plurality of medical data examples, the plurality of medical data examples ranked according to their degree of severity of disease
Claim 20:
receiving the reference image set;
performing pairwise comparison of each image of the plurality of images within the reference image set against every other image of the plurality of images within the image set; and
ranking the plurality of example fundus images according to a degree of severity of disease based on the pairwise comparisons
Claim 22:
receiving the reference data set; and
performing pairwise comparison of each medical data example within the reference data set against every medical data example within the reference data set; and
ranking the plurality of medical data examples according to a degree of severity of disease based on the pairwise comparisons
The difference between the present application and App# 18/994,127 is that the present application discloses the data being fundus image(s) which is obvious over Solanki et al. (US20150110368A1) ([0112], [0046], [0120], [0343], and [0346]) with the motivation for timely treatment and better quality of life (See Solanki, Background) in view of Burlina et al. (US20150265144A1) ([0045) with the motivation for detecting and classifying the severity of retinal disease (See Burlina, Background) and further in view of Nozaki et al. (US11024031B1) ([pg. 3]) with the motivation to determine severity using pairwise comparison.
Dependent claim 2 is an obvious variant of claim 2 of App# 18/994,127 for recitation of determining a position of the fundus image against the example fundus images within the reference image set; and comparing the position of the fundus image with a threshold for referral, wherein the flagging the fundus image as requiring referral comprises flagging the fundus image as requiring referral if the position of the fundus image is above the threshold for referral. Dependent claim 5 is an obvious variant of claim 5 of App# 18/994,127 for reciting wherein the one or more convolutional neural networks is a plurality of convolutional neural networks. Dependent claim 6 is an obvious variant of claim 6 of App# 18/994,127 for reciting wherein each of the plurality of convolutional neural networks is trained on a different data set. Dependent claim 7 is an obvious variant of claim 7 of App# 18/994,127 for reciting amalgamating the results of the pairwise comparisons from each convolutional neural network, and wherein the sending the fundus image for referral for investigation for the disease is based on the amalgamated pairwise comparisons. Dependent claim 8 is an obvious variant of claim 8 of App# 18/994,127 for reciting wherein the amalgamating of the results of the pairwise comparisons comprises supplying the results of the pairwise comparison as inputs to one or more lasso regression models, the or each lasso regression model having a decision boundary associated with a threshold of disease severity. Dependent claim 9 is an obvious variant of claim 9 of App# 18/994,127 for reciting wherein the or each lasso regression model is a plurality of lasso regression models and wherein the threshold of disease severity for each model is different. Dependent claim 10 is an obvious variant of claim 10 of App# 18/994,127 for reciting performing bootstrapping to estimate a certainty value of a probability of needing referral. The remaining dependent claims in the present application also recite substantially similar limitations to App 18/994,127 such as comparing accumulated results to the threshold for referral, determining a frequency of occurrence, thresholds corresponding to one of the plurality of thresholds for disease severity, performing linear discriminant analysis, using a selecting algorithm, and using an ensemble process.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-2, 5-10 and 12-23 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Step 1
The claim(s) recite(s) subject matter within a statutory category as a process (claim 1-2, 5-10, 12-19, and 22-23), process (claim 20), and article of manufacture (claim 21).
INDEPENDENT CLAIMS
Step 2A Prong 1
Claim 1 recites steps of
performing a pairwise comparison of the fundus image against each example fundus image in a reference image set using one or more convolution neural networks to determine a probability or a decision as to whether the fundus image is more or less severe than each example fundus image in the reference image set, and generating a comparison output representing the probability or the decision for each pairwise comparison; and
flagging the fundus image as requiring referral for investigation for the disease based on results of the pairwise comparisons by mapping the generated comparison outputs to an ordered severity scale of the reference image set,
wherein the reference image set includes a plurality of example fundus images, the plurality of example fundus images ranked according to their degree of severity of disease.
Claim 20 recites steps of
receiving the reference image set;
performing pairwise comparison of each image of the plurality of example fundus images within the reference image set against every other image of the plurality of example fundus images within the reference image set to determine a probability or a decision as to whether each image is more or less severe than every other image, and generating a comparison output representing the probability or the decision for each pairwise comparison; and
ranking the plurality of example fundus images according to a degree of severity of disease based on the comparison outputs from the pairwise comparisons.
These steps for determining if a fundus image requires referral for investigation for a disease, as drafted, under the broadest reasonable interpretation, includes performance of the limitations in the mind. That is, nothing in the claim element precludes the italicized portions from practically being performed in the mind through the evaluation and determination of a degree of severity of disease based on the pairwise comparisons. This could be analogized to collecting information, analyzing it, and displaying certain results of the collection and analysis (MPEP 2106). If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitations in the mind but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
Step 2A Prong 2
This judicial exception is not integrated into a practical application. In particular, the additional elements, non-italicized portions identified above for claims 1 and 20, do not integrate the abstract idea into a practical application, other than the abstract idea per se, because the additional elements amount to no more than limitations which:
amount to mere instructions to apply an exception (such as recitation of using one or more convolution neural networks amounts to invoking computers as a tool to perform the abstract idea, see MPEP 2106.05(f))
add insignificant extra-solution activity to the abstract idea (such as recitation of receiving the reference image set amounts to mere data gathering since it does not add meaningful limitations to the receiving action performed, see MPEP 2106.05(g))
Each of the above additional elements therefore only amounts to mere instructions to implement functions within the abstract idea using generic computer components or other machines within their ordinary capacity and add insignificant extra-solution activity to the abstract idea. There is no indication that the combination of elements improves the functioning of a computer or improves any other technology. These elements are therefore not sufficient to integrate the abstract idea into a practical application. Therefore, the above claims, as a whole, are directed to an abstract idea.
Step 2B
The claim(s) do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to discussion of integration of the abstract idea into a practical application, the additional elements amount to no more than mere instructions to apply an exception and add insignificant extra-solution activity to the abstract idea. Additionally, the additional limitations, other than the abstract idea per se, amount to no more than limitations which:
amount to mere instructions to apply an exception in particular fields such as an using one or more convolution neural networks, e.g., requiring the use of software to tailor information and provide it to the user on a generic computer, see Intellectual Ventures I LLC v. Capital One Bank, MPEP 2106.05(f)
amount to elements that have been recognized as well-understood, routine, and conventional activity in particular fields such as recitation of receiving the reference image set amounts; e.g., receiving or transmitting data over a network, Symantec, MPEP 2106.05(d)(II)(i).
DEPENDENT CLAIMS
Step 2A Prong 1
Dependent claims recite additional subject matter which further narrows or defines the abstract idea embodied in the claims (such as claims 2, 5-10, 12-19, and 21-23 reciting particular aspects for determining if a fundus image requires referral for investigation for a disease such as
[Claim 2] determining a position of the fundus image against the example fundus images within the reference image set; and
comparing the position of the fundus image with a threshold for referral,
wherein the flagging the fundus image as requiring referral comprises flagging the fundus image as requiring referral if the position of the fundus image is above the threshold for referral;
[Claim 5] wherein the one or more convolutional neural networks is a plurality of convolutional neural networks;
[Claim 6] wherein each of the plurality of convolutional neural networks is trained on a different data set;
[Claim 7] amalgamating the results of the pairwise comparisons from each convolutional neural network, and wherein the sending the fundus image for referral for investigation for the disease is based on the amalgamated pairwise comparisons;
[Claim 8] wherein the amalgamating of the results of the pairwise comparisons comprises supplying the results of the pairwise comparison as inputs to one or more lasso regression models, the or each lasso regression model having a decision boundary associated with a threshold of disease severity;
[Claim 9] wherein the or each lasso regression model is a plurality of lasso regression models and wherein the threshold of disease severity for each model is different;
[Claim 10] performing bootstrapping to estimate a certainty value of a probability of needing referral;
[Claim 12] wherein the amalgamating the results of the pairwise comparison comprises accumulating results from each convolutional neural network, and comparing the accumulated results to the threshold for referral;
[Claim 13] wherein the amalgamating the results comprises setting a plurality of thresholds of disease severity within the reference image set, and determining a frequency of occurrence of the fundus image above each of the thresholds of disease severity;
[Claim 14] wherein the threshold for referral corresponds to one of the plurality of thresholds of disease severity;
[Claim 15] wherein the amalgamating the results of the pairwise comparison comprises fitting an s-curve to the results of each convolutional neural network; determining a probability of requiring referral based each fitted s-curve; and performing linear discriminant analysis on the determined probabilities;
[Claim 16] comprising selecting a subset of the plurality of convolutional neural networks using a selecting algorithm;
[Claim 17] wherein the selecting algorithm comprises a lasso regression model, the lasso regression model having a decision boundary associated with the threshold for referral, the selecting comprising applying the results from each convolutional neural network into the lasso regression model, ordering the convolutional neural networks in terms of accuracy at predicting referral, and selecting a predetermined number of the highest ranked convolutional neural networks as the subset;
[Claim 18] receiving the reference image set;
performing pairwise comparison of each image of the plurality of example fundus images within the reference image set against every other image of the plurality of example fundus images within the image set; and
ranking the plurality of example fundus images according to a degree of severity of disease based on the pairwise comparisons;
[Claim 21] A non-transitory computer-readable medium including instructions stored thereon that when executed by one or more processors cause the processor to perform the method of claim 1;
[Claim 22] supplying the fundus image as an input vector of grayscale values in an RGB format,
wherein the grayscale values are concatenated together to form the input vector; and supplying each example fundus image in the reference image set as a reference input vector of grayscale values in the RGB format;
[Claim 23] extracting, via encoders of the one or more convolutional neural networks, a patient feature vector from the fundus image and a reference feature vector from each compared example fundus image,
wherein the grayscale values are concatenated together to form the input vector; and supplying each example fundus image in the reference image set as a reference input vector of grayscale values in the RGB format;
these italicized portions are mental since they merely describe types of data and determinations that can be performed by humans. In addition, then claims also recite mathematical calculations such as s-curve fitting, bootstrapping, vectors, concatenation, and linear discriminant analysis. The italicized portions containing the recitation of the training step has also been treated as part of the abstract idea, specifically as mathematical calculations which falls within the abstract idea of mathematical concepts, in light of the 2024 USPTO AI Guidance
Step 2A Prong 2
Dependent claims 5-9, 12, 15-18, 21, and 23 recite additional subject matter which amount to limitations consistent with the additional elements in the independent claims (the additional limitations in claim 5 (wherein the one or more convolutional neural networks is a plurality of convolutional neural networks); claim 6 (convolutional neural networks); claim 7 (each convolutional neural network); claim 8 (one or more lasso regression models, the or each lasso regression model); claim 9 (wherein the or each lasso regression model is a plurality of lasso regression models); claim 12 (each convolutional neural network); claim 15 (each convolutional neural network); claim 16 (the plurality of convolutional neural networks using a selecting algorithm); claim 17 (wherein the selecting algorithm comprises a lasso regression model, the lasso regression model; applying the results from each convolutional neural network into the lasso regression model; and, the convolutional neural networks); claim 21 (A non-transitory computer-readable medium including instructions stored thereon that when executed by one or more processors cause the processor); and, claim 23 (via encoders of the one or more convolutional neural networks) amounts to invoking computers as a tool to perform the abstract idea, see MPEP 2106.05(f)); and, add insignificant extra-solution activity to the abstract idea (such as recitation of claim 18 (receiving the reference image set); and, claim 23 (extracting […] a patient feature vector from the fundus image and a reference feature vector from each compared example fundus image) amounts to mere data gathering since it does not add meaningful limitations to the receiving action performed, see MPEP 2106.05(g))). Accordingly, these additional elements do not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea.
Step 2B
Dependent claims 3-9, 12, 15-17, 21, and 23 recite additional subject matter which, as discussed above with respect to integration of the abstract idea into a practical application, amount to invoking computers as a tool to perform the abstract idea, e.g., a commonplace business method or mathematical algorithm being applied on a general-purpose computer, Alice Corp. v. CLS Bank, MPEP 2106.05(f). There is no indication that these additional elements improve the functioning of a computer or improves any other technology. Their collective functions merely provide generic computer implementation. Dependent claims 18 and 23 recite additional subject matter which amount to elements that have been recognized as well-understood, routine, and conventional activity in particular fields such as recitation of receiving the reference image set amounts; e.g., receiving or transmitting data over a network, Symantec, MPEP 2106.05(d)(II)(i).
Therefore, in consideration of all the facts, the present invention is not a patent-eligible invention under USC 101. Additionally, it is evident that the present claims monopolize every possible application of pairwise comparison to rank fundus images, restricting further innovation in this area without offering a specific, technical improvement to how the computer actually operates. Improved predictive accuracy or using AI tools is generally not enough to transform an abstract idea into patent-eligible subject matter if the core of the invention is still a method of determination; “monopolization of those tools through the grant of a patent might tend to impede innovation more than it would tend to promote it.” Alice Corp., 573 U.S. at 216, 110 USPQ2d at 1980 (quoting Myriad, 569 U.S. at 589, 106 USPQ2d at 1978 and Mayo Collaborative Servs. v. Prometheus Labs. Inc., 566 U.S. 66, 71, 101 USPQ2d 1961, 1965 (2012)).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35
U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 148
USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-2, 5-7, 10-15, and 18-23 are rejected under 35 U.S.C. 103 as being unpatentable over Solanki et al. (US20150110368A1) in view of Burlina et al. (US20150265144A1) and further in view of Nozaki et al. (US11024031B1).
Regarding claim 1, Solanki discloses […] the fundus image […] ([0112] “automatically evaluate fundus photographs”)
and generating a comparison output representing the probability or the decision for each […] ([0343] “the probability that a blob is classified as an MA or the probability that two blobs are marked as matched MAs and hence persistent is estimated”)
and flagging the fundus image as requiring referral for investigation for the disease based on results […] ([0246] “whereas images with hazy media are flagged as being of insufficient quality for effective grading. Quality assessment can allow the clinician to determine whether he needs to immediately reimage the eye or refer the patient to a clinician”)
wherein the reference image set includes a plurality of example fundus images ([0340] “In one embodiment, binary images Bb and Bl with lesions of interest marked out are created for the baseline and longitudinal images.”)
Solanki does not explicitly disclose however Burlina teaches the plurality of example fundus images ranked according to their degree of severity of disease ([0014] “FIG. 1 shows examples of four fundus images with increasing AMD severity. Upper left: an AMD category 1 (no AMD). Upper right: an AMD category 2 (early AMD); Lower left: an AMD category 3 (intermediate AMD) with geographic atrophy not involving the center of the retina; Lower right: an AMD category 4 with evidence of both neovascularization and geographic atrophy (advanced AMD)”)
by mapping the generated comparison outputs to an ordered
severity scale of the reference image set ([0045] “In some embodiments, the method can further include determining a likelihood of a severity of the retinal disease based on the comparison of the number of occurrences of each of the key image features in the retinal image of the individual to the reference data. Although some examples described in this specification use a specific number of severity levels, such as four, the general concepts of the current invention are not limited to a particular number of severity levels. There could just be one level in some embodiments, i.e., detecting only, or there could be two, three, four, or even more than four severity levels in other embodiments.”)
It would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki the plurality of example fundus images ranked according to their degree of severity of disease; and, mapping the generated comparison outputs to an ordered severity scale of the reference image set as taught by Burlina since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Solanki in view of Burlina does not explicitly disclose however Nozaki teaches performing a pairwise comparison of the […] image against each example […] image in a reference image set using one or more convolution neural networks to determine a probability or a decision as to whether the […] image is more or less severe than each example […] image in the reference image set; pairwise comparison; of the pairwise comparisons ([pg. 3] “The method for determining a severity of gastric cancer in a subject may include obtaining various images of a stomach of the subject including wavelength images, and generating difference images from the wavelength images. The subject images may then be compared with reference images representative of different severity levels of gastric cancer, or input into a learned model trained using the reference images stored in the database to extract a feature pattern corresponding to a severity of gastric cancer to diagnose the subject as having a particular severity level of gastric cancer.”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and the detection and classification of Burlina performing a pairwise comparison of the […] image against each example […] image in a reference image set using one or more convolution neural networks to determine a probability or a decision as to whether the […] image is more or less severe than each example […] image in the reference image set as taught by Nozaki since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 2, Solanki discloses determining a position of the fundus image against the example fundus images within the reference image set; and comparing the position of the fundus image with a threshold for referral ([0016] “identifying pixels in the first image as landmarks; identifying pixels in the second image as landmarks; computing descriptors at landmark pixels; matching descriptors across the first image and the second image; and estimating a transformation model to align the first image and the second image; compute changes in lesions and anatomical structures in registered images; and quantify the changes in terms of statistics” [0392] “DR screening block 114 determines whether a particular fundus image includes abnormalities indicative of diabetic retinopathy such that the patient should be referred to an expert”)
wherein the flagging the fundus image as requiring referral comprises flagging the fundus image as requiring referral if the position of the fundus image is above the threshold for referral ([0022] “In one embodiment of the system, this lesion information is further processed to generate a prediction score indicating the severity of retinopathy in the patient, which provides context determining potential further operations such as clinical referral or recommendations for the next screening date.”)
Regarding claim 5, Solanki discloses wherein the one or more convolutional neural networks is a plurality of convolutional neural networks ([0387] “In one embodiment, the system uses convolution networks, sometimes referred to as cony-nets, based classifiers, which are deep architectures that have been shown to generalize well for visual inputs.”)
Regarding claim 6, Solanki discloses wherein each of the plurality of convolutional neural networks is trained on a different data set ([0120] “using algorithms that learn, directly or indirectly, from a set of examples of already graded fundus images. This training data could have a key influence on the sensitivity and specificity of the algorithm. […] in some embodiments, the computerized screening algorithm performs well on cross-dataset testing, that is, the algorithm generalizes well, when trained on one dataset and tested on another.”)
Regarding claim 7, Solanki discloses amalgamating the results […] from each convolutional neural network, and wherein the sending the fundus image for referral for investigation for the disease is based on the amalgamated […] ([0021] “using the descriptors to classify image suitability for grading comprising one or more of: […] convolution networks” [0022] “A joint segmentation-recognition method can be used to recognize and localize the lesions and retinal structures. In one embodiment of the system, this lesion information is further processed to generate a prediction score indicating the severity of retinopathy in the patient, which provides context determining potential further operations such as clinical referral or recommendations for the next screening date.”)
Solanki in view of Burlina does not explicitly disclose however Nozaki teaches […] of the pairwise comparisons […] ([pg. 3] “The subject images may then be compared with reference images representative of different severity levels” [pg. 20] “facilitate such comparisons”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and the detection and classification of Burlina a pairwise comparison as taught by Nozaki since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 10, Solanki discloses performing bootstrapping to estimate a certainty value of a probability of needing referral ([0359] “Classifying encounters using encounter-level descriptors computed in (f) as normal or abnormal (one classifier each for each abnormality, lesion, or disease) using one or more of supervised learning techniques including but not limited to: […] (viii) convolution networks, […] with […] bootstrap aggregation.”)
Regarding claim 12, Solanki discloses wherein the amalgamating the results […] comprises accumulating results from each convolutional neural network, and comparing the accumulated results to the threshold for referral ([0013] “normal or abnormal using supervised learning utilizing the computed descriptors, using one or more of: […] convolution networks [0335] “The results of the analysis are provided to the mobile device 289008, which performs an initial interpretation of the results and presents a diagnosis report to the patient.” [0094] “FIGS. 41C and 41D show embodiments of Cytomegalovirus retinitis screening results using one embodiment of the Cytomegalovirus retinitis detection module for “retina with CMVR” category screened as “refer”.”)
Solanki in view of Burlina does not explicitly disclose however Nozaki teaches […] of the pairwise comparison […] ([pg. 3] “The subject images may then be compared with reference images representative of different severity levels”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and the detection and classification of Burlina […] of the pairwise comparisons […]; […] of the pairwise comparison […] as taught by Nozaki since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 13, Solanki does not explicitly disclose however Burlina teaches wherein the amalgamating the results comprises setting a plurality of thresholds of disease severity within the reference image set ([0091] “A set of fundus photographs were digitized from 595 of these 600 patients forming a set of over 72,000 images graded for AMD severity.” [0092] “Specifically, each image is assigned an AMD category from 1 to 4, with 1 representing images showing minimal to no evidence of AMD; category 2 corresponding to the early stage of AMD25; category 3 corresponding to the intermediate stage of AMD; and category 4 representing images from patients with the advanced stage of AMD.”)
and determining a frequency of occurrence of the fundus image above each of the thresholds of disease severity [0095]
PNG
media_image1.png
241
420
media_image1.png
Greyscale
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki wherein the amalgamating the results comprises setting a plurality of thresholds of disease severity within the reference image set, and determining a frequency of occurrence of the fundus image above each of the thresholds of disease severity as taught by Burlina since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 14, Solanki does not explicitly disclose however Burlina teaches wherein the threshold for referral corresponds to one of the plurality of thresholds of disease severity ([0014] “[…] (no AMD)[…] (early AMD) […] (intermediate AMD) […] (advanced AMD)”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki wherein the threshold for referral corresponds to one of the plurality of thresholds of disease severity as taught by Burlina since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 15, Solanki discloses wherein the amalgamating the results […] comprises fitting an s-curve to the results of each convolutional neural network ([0387] “In one embodiment, the system uses convolution networks” [0405] “The ROC curves for this example implementation, shown in FIG. 40, demonstrate an impressive cross-dataset testing performance, especially for detecting DR onset (AUROC of 0.91).”)
determining a probability of requiring referral based each fitted s-curve ([0325] “the class probability for lesion class and non-lesion class are computed and are used as the decision statistic.”)
and performing linear discriminant analysis on the determined probabilities ([0325] “A decision statistic for a particular pixel is generated by computing the distance of the given element from the given lesion classification hyper plane defined […] using Fisher linear discriminant or the like.”)
Solanki in view of Burlina does not explicitly disclose however Nozaki teaches […] of the pairwise comparison […] ([pg. 3] “The subject images may then be compared with reference images representative of different severity levels” [pg. 20] “facilitate such comparisons”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and the detection and classification of Burlina […] of the pairwise comparison […] as taught by Nozaki since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 18, Solanki discloses receiving the reference image set ([0244] “On a dataset of 10,104 images with over 2000 lens shot images on 50-50 train-test splits area under receiver operating characteristics (ROC) curve (AUROC) of over 0.998 were obtained.”)
performing […] of each image of the plurality of images within the reference image set against every other image of the plurality of example fundus images within the image set ([0337] “The first image in the longitudinal series is referred to as the baseline image Ib and any other registered longitudinal image is denoted as I1.” [0341] “In one embodiment, these pixels are then classified using lesion classifier […] Newly appearing lesions can be found by labeling image I and comparing those regions to corresponding regions in Ib to identify newly appearing lesions.”)
Solanki does not explicitly disclose however Burlina teaches and
ranking the plurality of example fundus images according to a degree of severity of disease […] ([0014] “FIG. 1 shows examples of four fundus images with increasing AMD severity. Upper left: an AMD category 1 (no AMD). Upper right: an AMD category 2 (early AMD); Lower left: an AMD category 3 (intermediate AMD) with geographic atrophy not involving the center of the retina; Lower right: an AMD category 4 with evidence of both neovascularization and geographic atrophy (advanced AMD)”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki ranking the plurality of example fundus images according to a degree of severity of disease based on the pairwise comparisons as taught by Burlina since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Solanki in view of Burlina does not explicitly disclose however Nozaki teaches […] pairwise comparison […]; […] based on the pairwise comparisons […] ([pg. 3] “The subject images may then be compared with reference images representative of different severity levels” [pg. 20] “facilitate such comparisons”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and the detection and classification of Burlina […] pairwise comparison […]; […] based on the pairwise comparisons […] as taught by Nozaki since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 19, Solanki discloses wherein the disease is diabetic retinopathy ([0105] “Methods and systems are disclosed that provide automated image analysis allowing detection, screening, and/or monitoring of retinal abnormalities, including diabetic retinopathy”)
Regarding claim 20, Solanki discloses receiving the reference image set ([0244] “On a dataset of 10,104 images with over 2000 lens shot images on 50-50 train-test splits area under receiver operating characteristics (ROC) curve (AUROC) of over 0.998 were obtained.”)
[…] the fundus image […] ([0112] “automatically evaluate fundus photographs”)
and generating a comparison output representing the probability or the decision for each […] ([0343] “the probability that a blob is classified as an MA or the probability that two blobs are marked as matched MAs and hence persistent is estimated”)
Solanki does not explicitly disclose however Burlina teaches and
ranking the plurality of example fundus images according to a degree of severity of disease based on the comparison outputs […]([0094] “classify fundus images into different levels of AMD severity” [0014] “FIG. 1 shows examples of four fundus images with increasing AMD severity. Upper left: an AMD category 1 (no AMD). Upper right: an AMD category 2 (early AMD); Lower left: an AMD category 3 (intermediate AMD) with geographic atrophy not involving the center of the retina; Lower right: an AMD category 4 with evidence of both neovascularization and geographic atrophy (advanced AMD)”)
It would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki ranking the plurality of example fundus images according to a degree of severity of disease based on the pairwise comparisons as taught by Burlina since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Solanki in view of Burlina does not explicitly disclose however Nozaki teaches performing pairwise comparison of each image of the plurality of example […] images within the reference image set against every other image of the plurality of example […] images within the reference image set to determine a probability or a decision as to whether each image is more or less severe than every other image; pairwise comparison; from the pairwise comparisons ([pg. 3] “The method for determining a severity of gastric cancer in a subject may include obtaining various images of a stomach of the subject including wavelength images, and generating difference images from the wavelength images. The subject images may then be compared with reference images representative of different severity levels of gastric cancer, or input into a learned model trained using the reference images stored in the database to extract a feature pattern corresponding to a severity of gastric cancer to diagnose the subject as having a particular severity level of gastric cancer.”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and the detection and classification of Burlina performing pairwise comparison of each image of the plurality of example […] images within the reference image set against every other image of the plurality of example […] images within the reference image set to determine a probability or a decision as to whether each image is more or less severe than every other image; pairwise comparison; from the pairwise comparisons as taught by Nozaki since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 21, Solanki discloses A non-transitory computer-readable medium including instructions stored thereon that when executed by one or more processors cause the processor to perform the method of claim 1 ([0021] “In another embodiment, non-transitory computer storage that stores executable program instructions is disclosed.”)
Regarding claim 22, Solanki discloses supplying the fundus image as an input vector of grayscale values in an RGB format, wherein the grayscale values are concatenated together to form the input vector ([0248] “FIG. 18 gives an overview of one embodiment of a process for performing image quality computation.” [0262] “In one embodiment, normalized RGB color histograms are computed over the whole image and used as descriptors of color. In one embodiment, the computed descriptor has a length of 20 for each of the R, G, and B channels.” [0293] “obtain a number or a vector of numbers for each scale (“multi-scale analysis”), which are then concatenated to make a composite vector of numbers (“multi-scale descriptors”)”)
supplying each example fundus image in the reference image set as a reference input vector of grayscale values in the RGB format ([0337] “In one embodiment, longitudinal retinal fundus images are registered to the baseline image as described in the section above entitled “Image Registration”.” [0251] “In one embodiment, using 3-channel (RGB) color retinal fundus images”)
Regarding claim 23, Solanki discloses […] a patient feature vector from the fundus image […] compared example fundus image; and mathematically combining the patient feature vector and the reference feature vector using a merging layer of the one or more convolutional neural networks to generate a merged feature vector ([0112] “automatically evaluate fundus photographs”) [0387] “In one embodiment, the system uses convolution networks” [0189] “In another embodiment, the number or the vectors of numbers can be computed on a multitude of images obtained by progressively scaling […] down the original input image with a fixed scaling factor referred to as multi-scale analysis, and stacking the obtained vector of numbers into a single larger vector of numbers”)
Solanki in view of Burlina does not explicitly disclose however Nozaki teaches extracting, via encoders of the one or more convolutional neural networks, […] and a reference feature vector from each ([pg. 9] “For instance, an abnormal image may be obtained and compressed with the encoder.” [pg. 32] “The reference images are then vectorized in step 1106.”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and the detection and classification of Burlina extracting, via encoders of the one or more convolutional neural networks, […] and a reference feature vector from each as taught by Nozaki since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Claims 8 and 9 are rejected under 35 U.S.C. 103 as being unpatentable over Solanki et al. (US20150110368A1) in view of Burlina et al. (US20150265144A1), Nozaki et al. (US11024031B1), and further in view of Oh et al. (Diabetic retinopathy risk prediction for fundus examination using sparse learning: a cross-sectional study).
Regarding claim 8, Solanki in view of Burlina does not explicitly disclose however Nozaki teaches […] of the pairwise comparisons […]; […] of the pairwise comparison […] ([pg. 3] “The subject images may then be compared with reference images representative of different severity levels” [pg. 20] “facilitate such comparisons”)
It would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and the detection and classification of Burlina […] of the pairwise comparisons […]; […] of the pairwise comparison […] as taught by Nozaki since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Solanki in view of Burlina and Nozaki does not explicitly disclose however Oh teaches wherein the amalgamating of the results […] comprises supplying the results […] as inputs to one or more lasso regression models ([pg. 5] “When the optimal values of λ were applied for training the penalized logistic regression including […] LASSO, the resulting coefficients models that we obtained are given in Additional file 1. The coefficients of OLR and LR-BS were also calculated.”)
the or each lasso regression model having a decision boundary associated with a threshold of disease severity ([pg. 5] “We generated the ROC curves and selected cut-off points that maximized Youden's index [32]. Participants above the cut-off points were classified as being at high risk in each prediction model.”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki, retinal disease detection system of Burlina, and diagnosis system of Nozaki wherein the amalgamating of the results of the pairwise comparisons comprises supplying the results of the pairwise comparison as inputs to one or more lasso regression models, the or each lasso regression model having a decision boundary associated with a threshold of disease severity as taught by Oh since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Regarding claim 9, Solanki in view of Burlina and Nozaki does not explicitly disclose however Oh teaches wherein the or each lasso regression model is a plurality of lasso regression models and wherein the threshold of disease severity for each model is different ([pg. 4] “The form of logistic regression was used for all prediction models due to dichotomous clinical outcome […]
PNG
media_image2.png
127
332
media_image2.png
Greyscale
[pg. 5] “selected cut-off points that maximized Youden's index [32].”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki, retinal disease detection system of Burlina, and diagnosis system of Nozaki wherein the or each lasso regression model is a plurality of lasso regression models and wherein the threshold of disease severity for each model is different as taught by Oh since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Claims 16 is rejected under 35 U.S.C. 103 as being unpatentable over Solanki et al. (US20150110368A1) in view of Burlina et al. (US20150265144A1), Nozaki et al. (US11024031B1), and further in view of Xia et al. (CN112686902A).
Regarding claim 16, Solanki in view of Burlina and Nozaki does not explicitly disclose however Xia teaches selecting a subset of the plurality of convolutional neural networks using a selecting algorithm ([pg. 4] “FCNN Representing 128-dimensional convolutional neural network features, Fradiomics Representing the image omics characteristics on four sequences of 4 × 104 dimensions, F comprises 544-dimensional characteristics in total; step 5.4: l1 regularization feature selection, wherein feature selection is performed by adopting an L1 regularization Lasso algorithm (Lasso)”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and retinal disease, detection system of Burlina, and diagnosis system of Nozaki selecting a subset of the plurality of convolutional neural networks using a selecting algorithm as taught by Xia since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Claims 17 is rejected under 35 U.S.C. 103 as being unpatentable over Solanki et al. (US20150110368A1) in view of Burlina et al. (US20150265144A1), Nozaki et al. (US11024031B1), Xia et al. (CN112686902A), and further in view of Bu et al. (Ensemble of Deep Convolutional Learning Classifier System Based on Genetic Algorithm for Database Intrusion Detection).
Regarding claim 17, Solanki in view of Burlina and Nozaki does not explicitly disclose however Xia teaches wherein the selecting algorithm comprises a lasso regression model ([pg. 4] “wherein feature selection is performed by adopting an L1 regularization Lasso algorithm (Lasso)”)
the lasso regression model having a decision boundary associated with the threshold for referral ([pg. 5] “according to the classification mark obtained by each pixel point, summarizing the pixel points (x, y) obtaining the same classification result to form the tumor boundary and the labels of different tumor tissues […] determining the classification label of the region so as to determine the label of the central pixel of the region”)
the selecting comprising applying the results from each convolutional neural network into the lasso regression model ([pg. 4] “FCNN Representing 128-dimensional convolutional neural network features, Fradiomics Representing the image omics characteristics on four sequences of 4 × 104 dimensions, F comprises 544-dimensional characteristics in total; step 5.4: l1 regularization feature selection, wherein feature selection is performed by adopting an L1 regularization Lasso algorithm (Lasso)”)
It would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki and retinal disease, detection system of Burlina, and diagnosis system of Nozaki wherein the selecting algorithm comprises a lasso regression model, the lasso regression model having a decision boundary associated with the threshold for referral, the selecting comprising applying the results from each convolutional neural network into the lasso regression model as taught by Xia since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Solanki in view of Burlina, Nozaki, and Xia does not explicitly disclose however Bu teaches ordering the convolutional neural networks in terms of accuracy at predicting referral ([pg. 3] “It is known that the intrusion detection system accuracy can be enhanced by utilizing the ensemble of multiple models [25,26]. […] finally, the ensemble collection component uses the results of each generated models to select and combine relevant models into an ensemble with respect to accuracy and diversity”)
and selecting a predetermined number of the highest ranked convolutional neural networks as the subset ([pg. 9] “Afterward, starting with the model with the highest accuracy, a set of models are collected such that Qav of the ensemble is reduced until the number of models within the ensemble is more or equal to the minimum number of required models, and no other models can be added to the set without increasing the ensemble’s Qav […] The details of the ensemble collection process for selecting the heterogeneous model considering diversity are described in Algorithm 1.
S ← {}, P ← Collection of generated models;
x ← best accuracy model in P;”)
Therefore, it would have obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to include in the screening and diagnosis system of Solanki, retinal disease detection system of Burlina, diagnosis system of Nozaki, and the two-stage calculation system of Xu ordering the convolutional neural networks in terms of accuracy at predicting referral, and selecting a predetermined number of the highest ranked convolutional neural networks as the subset as taught by Bu since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art.
Response to Arguments
Applicant’s arguments filed on 03 August 2026 have been considered but are not fully persuasive.
Regarding the double patenting rejection, applicant argues that amended claim 1 performs pairwise comparison of a patient's fundus image against each example fundus image in a reference image set (representing different patients/severities) ranked according to their degree of severity of disease while Solanki completely fails to disclose or suggest a pre-ranked reference set of different example images that acts as a calibrated "ruler" spanning a spectrum of disease severities. Applicant further asserts that the claims of US ‘247 merely recite using generic "machine learning algorithms" to determine a "difference in severity" in comparison to amended claim 1. Additionally, applicant asserts that neither Solanki nor the US ‘247 disclose “mapping the generated comparison outputs to an ordered severity scale of the reference image set.” That the claims of US '247 merely flag an image "based on results of the pairwise comparisons" at a generic level, and Solanki has no ordered severity scale (no "ruler" database) and therefore cannot map comparison outputs onto such a scale. Applicant asserts that it would not be obvious to combine Solanki and US ‘247 as this would take significant inventive effort and would not be obvious.
Examiner disagrees with the applicant’s arguments. Examiner asserts the amendments do not move the needle on the double patenting rejection. Applicant needs more significant amendments to advance prosecution. Solanki is a narrow interpretation of the applicant’s generic claims. Nowhere in the applicant’s claims is there any recitation that the reference image set represents different patients/severities. Examiner will not take such a narrow interpretation. Examiner is also confused on where applicant is getting 18/991,247 from. The subject matter of that application does not match the one cited by the examiner. Regardless, the applicant’s amendments also do little to distinguish from the Zheng application since all the applicant has done is try to specify terms that are broader in the Zheng application. The disclosure of the cited Zheng application is the exact same disclosure as the present application; even pairwise comparison is specifically disclosed contrary to the applicant’s assertion. A person having ordinary skill in the art would easily combine the claims of the Zheng application with Solanki. It’s unclear how this would take significant inventive effort. Therefore, the double patenting rejection has been maintained.
Regarding the USC 101 rejection, applicant argues that the claims do not recite an abstract idea since claim 1 recites a complex process performed by convolution neural networks and thus by the terms of itself cannot be performed in the human mind. Applicant asserts that a CNN is a specific, multi-layered computational architecture implemented in computer hardware ([0048] of specification) and it uses pixel data through mathematical convolution, pooling, and activation operations that cannot be duplicated or performed in the human mind. That a human mind cannot ingest raw pixel arrays, execute parallel weight-sharing matrix multiplications, and output mathematically calibrated probabilities or binary decisions for a series of dozens or hundreds of image comparisons. Applicant states that since the claim is structurally anchored to the execution of hardware-constrained convolutional neural networks operating on digital image vectors, it does not recite a mental process or any other category of abstract idea. For Step 2A Prong 2, applicant submits that the additional elements integrate the exception into a practical application by providing a specific technological improvement in medical image processing. Applicant cites multiple paragraphs of the specification asserting the amended claims address the problems faced by conventional methods of "insufficient accuracy and flexibility" (i) performing pairwise comparisons of a patient's fundus image against a pre-sorted reference set using convolutional neural networks to determine a relative probability or decision of severity; (ii) generating a discrete comparison output representing this relative determination for each pair; and (iii) mapping these comparison outputs to an ordered severity scale of the reference image set (the "ruler") to automate the referral decision. Applicant states that by shifting the computational task from direct image classification to a series of automated pairwise comparative steps mapped against a calibrated, pre-ranked reference scale, the claimed method achieves significantly higher accuracy and flexibility (paragraphs [0006], [0051] of the specification). Applicant asserts that amended claim 1 does not merely use a computer as a generic tool to execute a well-known administrative task but rather defines a specific mathematical and structural improvement in computer-aided diagnostics that increases the accuracy of ophthalmic disease detection which integrates any judicial exception into a practical application. Applicant further argues that amended claim 1 patent-eligible because the ordered combination of elements contributes to an inventive concept that is "significantly more" than the exception since the specification explicitly teaches that the claimed CNN architecture operates fundamentally differently from conventional image-comparison models ([0050] of specification). Applicant asserts that the claim recites an unconventional image-processing pipeline that goes far beyond generic, routine, or conventional computer implementations. Applicant requests withdrawal of the USC 101 rejection.
Examiner disagrees with the applicant’s arguments. Applicant’s claims are clearly directed to an abstract idea. The claimed limitations are tasks that human experts, such as ophthalmologists and retinal specialists, have done traditionally for decades, even the applicant’s specification admits that typically, a diagnosis is made by a trained medical professional ([0002] of specification). The MPEP makes it clear that claims can recite an abstract idea even if they are claimed as being performed on a computer. The courts have also found claims requiring a generic computer or nominally reciting a generic computer may still recite abstract idea even though the claim limitations are not performed entirely by a human. The limitations identified as abstract in the present application are very outcome-based or result-focused and don’t give much technical detail that goes beyond what a human can do. Merely adding a generic computer, generic computer components, or a programmed computer to perform generic computer functions does not automatically overcome an eligibility rejection. Alice Corp. Pty. Ltd. v. CLS Bank Int'l, 573 U.S. 208, 224, 110 USPQ2d 1976, 1984 (2014). See also OIP Techs. v. Amazon.com, 788 F.3d 1359, 1364, 115 USPQ2d 1090, 1093-94 (Fed. Cir. 2015) ("Just as Diehr could not save the claims in Alice, which were directed to ‘implement[ing] the abstract idea of intermediated settlement on a generic computer’, it cannot save OIP's claims directed to implementing the abstract idea of price optimization on a generic computer.") (citations omitted). Tacking on a CNN in the claim and asserting operations that are not claimed and not even disclosed in the specification (e.g., activation) does not overcome USC 101. The applicant is arguing under the presumption that simply claiming a CNN automatically means convolution, pooling, and activation operations, but examiner points out that one of ordinary skill in the art would recognize that a generic CNN claimed without additional detail could simply mean the combination of one Conv2D layer -> one Flatten layer -> one Dense layer; claim at present does not require what the applicant is arguing. Despite the applicant’s assertions, examiner asserts that the claim still puts no limit on how the computer actually performs the argued limitations (i.e., computer-focused operations) such they cannot be considered an abstract idea. Even if the claims nominally recite computer components that are rooted in technology, there is no recitation of how the computer components are specifically programmed to distinguish from generic computer processes. Similarly, the applicant is arguing that the CNN outputs mathematically calibrated probabilities or binary decisions for a series of dozens or hundreds of image comparisons, but examiner points out there is zero quantitative indication in the claim that would suggest dozens or hundreds of image comparisons; another example of applicant arguing under a presumption. Even with the applicant’s recitation of CNN, examiner further evaluated that recitation under Step 2A Prong 2. Claim 1 can reasonably be described as falling into the mental processes grouping of abstract ideas.
Under Step 2A Prong 2, the amendments do not do much to advance prosecution because the present specification provides a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art. The MPEP provides that improvements to the functioning of a computer or to any other technology or technical field can signal eligibility, see MPEP 2106.05(a), and provides examples of improvements to computer functionality, MPEP 2106.05(a)(I), and improvements to any other technology of technical field, MPEP 2106.05(a)(I). “In computer-related technologies, the examiner should determine whether the claim purports to improve computer capabilities or, instead, invokes computers merely as a tool”. Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1336, 118 USPQ2d 1684, 1689 (Fed. Cir. 2016). In Enfish, the court evaluated the patent eligibility of claims related to a self-referential database. Id. The court concluded the claims were not directed to an abstract idea, but rather to an improvement to computer functionality. Id. It was the specification' s discussion of the prior art and how the invention improved the way the computer stores and retrieves data in memory in combination with the specific data structure recited in the claims that demonstrated eligibility. 822 F.3d at 1339, 118 USPQ2d at 1691. The claim was not simply the addition of general-purpose computers added post-hoc to an abstract idea, but a specific implementation of a solution to a problem in the software arts. 822 F.3d at 1339, 118 USPQ2d at 1691. Unlike Enfish, the instant claimed invention appears to improve upon a judicial exception rather than a problem in the software arts. Rather than improving a computer's algorithm (i.e., solving a technically based problem), the claimed invention purports to solve the non-technological problem of insufficient medical personnel who are trained to assess the high number of fundus images from patients suspected as having an ocular disease such as diabetic retinopathy ([0002] of specification) by using computers/CNN as a tool to perform data processing/manipulation rather than a concrete technical improvement. In other words, one of the main/glaring issues with the present invention is that the problem solved by the applicant is not technological problem. This asserted improvement by the applicant is not an improvement to computer technology and is not considered patent eligible by the USPTO (see MPEP 2106). All the applicant is doing is applying known technology for their intended benefit(s) to a new data environment and calling it an improvement (see Customedia Techs., LLC v. Dish Network Corp., Case No.18-2239 (Fed. Cir. Mar. 6, 2020). Even if the claims provide a solution to a technical problem, examiner asserts that applicant is clearly using a standard, off-the-shelf machine learning primitive operating in its normal manner:
First, one of ordinary skill in the art would recognize that when the claim performs a pairwise comparison against each example image in the reference set where if the reference set scales to 1,000 baseline images to ensure clinical diversity, evaluating just one new patient image requires 1,000 separate Convolutional Neural Network (CNN) inferences. This is highly inefficient. Standard medical ML workflows use the CNN to extract features once and compare vector distances, or project the image directly onto a continuous latent space. The applicant’s specification does not explain/lacks detail on how exponential computational complexity is managed that is necessary for an improvement to be apparent to a person of ordinary skill in the art. Thus, it is not evident from the specification how the claims improve the operation of computers.
Second, one of ordinary skill in the art would recognize that relying on independent probabilistic decisions ("more or less severe") and mapping them back to an ordered scale leads to intransitivity and logic violations. Neural networks lack inherent logical transitivity. If the reference set is ordered (A < B < C), a CNN evaluating a new image [X] might confidently output that (X > B), (X < A), and (X > C). This logical loop (i.e., intransitivity) makes mapping the outputs to a strict linear scale highly volatile, prone to mathematical contradictions, and difficult to calibrate. The applicant’s specification does not explain/lacks detail on how intransitivity and ordering logic violations are accounted for that is necessary for an improvement to be apparent to a person of ordinary skill in the art. Thus, it is not evident from the specification how the claims improve the operation of computers.
Third, one of ordinary skill in the art would recognize that when an entire decision engine of the claim anchors directly to a specific, static reference image set, it leads to high sensitivity due to label noise and reference set drift. If a few reference images have unspotted labeling errors, artifacts, poor contrast, or variations due to different camera manufacturers, they will systematically skew every single subsequent patient diagnosis. Medical AI typically trains on global distributions to wash out this kind of local data noise. The applicant’s specification does not explain/lacks detail on how high sensitivity due to label noise and reference set drift are accounted for that is necessary for an improvement to be apparent to a person of ordinary skill in the art. Thus, it is not evident from the specification how the claims improve the operation of computers.
Fourth, one of ordinary skill in the art would recognize that forcing a binary or continuous severity comparison across independent images leads to confounding severity with image quality. Fundus photographs are notorious for technical variability (e.g., uneven illumination, cataracts, blur, or poor dilation). A CNN trained purely on severity comparisons may flag a highly blurry, mild-disease image as more severe than a crisp, clear baseline image simply because the network confuses poor image quality with pathological opacity. The applicant’s specification does not explain/lacks detail on how confounding severity with image quality is handled that is necessary for an improvement to be apparent to a person of ordinary skill in the art. Thus, it is not evident from the specification how the claims improve the operation of computers.
The examiner also asserts the following facts which the applicant will not be able to dispute:
1) the invention does NOT involve a novel algorithm or data structure that significantly improves the computer's functionality,
2) the invention does NOT involve a new hardware component or configuration that works with the computer to achieve a specific technical benefit, and
3) the computer is NOT used in a completely new way demonstrating a significant technical advancement.
It is evident from the specification and claims that the applicant is not improving computer technology, and instead providing an improvement to the abstract idea. An improvement to the abstract idea is not an improvement to computer technology. Thus, examiner does not see how the present claims improve the functioning of a computer or provide improvements to any other technology or technical field. The claimed invention appears similar to the example of improvements that are insufficient to show an improvement in computer-functionality such as arranging transactional information on a graphical user interface in a manner that assists traders in processing information more quickly, Trading Technologies v. IBG LLC, 921 F.3d 1084, 1093-94, 2019 USPQ2d 138290 (Fed. Cir. 2019). See MPEP 2106.05(a)(I)(viii). The broad claims are lacking concrete limitations to integrate the abstract idea into a practical application. Examiner points out that the claimed limitations have no indication in the specification that the operations recited invoke any inventive programming, require any specialized computer hardware or other inventive computer components, i.e., a particular machine, or that the claimed invention is implemented using other than generic computer components to perform generic computer functions. See DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1256 (fed Cir. 2014) (“[A]fter Alice, there can remain no doubt: recitation of generic computer limitations does not make an otherwise ineligible claim patent-eligible.”). Most importantly, in DDR Holdings & unlike the present claims, the claims at issue specified how interactions with the Internet were manipulated to yield a desired result—a result that overrode the routine and conventional sequence of events ordinarily triggered by the click of a hyperlink. 773 F.3d at 1258; 113 USPQ2d at 1106. The examiner also points out that there is no indication in the specification that the claimed invention affects a transformation or reduction of a particular article to a different state or thing. Examiner points to the recitation of convolution neural networks in the claim(s) as generic. "[T]he mere recitation of a generic computer cannot transform a patent-ineligible abstract idea into a patent-eligible invention." Alice Corp. v. CLS Banklnt'l, 573 U.S. 208 223 (2014). Applicant does not and cannot contend they invented the concept of convolution neural networks, nor does the specification disclose any new convolution neural network. The alleged improvement of using convolution
neural networks lies in the abstract idea itself, not to any technological improvement nor to any improvement to the functioning of a computer. See BSG Tech LLC v. Buyseasons, Inc., 899 F.3d 1281, 1287-88 (Fed. Cir. 2018). The fact pattern of the applicant’s claims is congruent to the Recentive Analytics, Inc. v. Fox Corp., 2025 U.S.P.Q.2d 628 (Fed. Cir. 2025) decision by the Federal Circuit. Just like in Recentive, the present claims do not delineate steps through which the convolution neural network achieves an improvement. See, e.g., IBM v. Zillow Grp., Inc., 50 F.4th 1371, 1381 (Fed. Cir. 2022) (holding abstract a claim that "d[id] not sufficiently describe how to achieve [its stated] results in a non-abstract way," because "[s]uch functional claim language, without more, is insufficient for patentability under our law." (quoting Two-Way Media Ltd v. Comcast Cable Commc'ns, LLC, 874 F.3d 1329, 1337 (Fed. Cir. 2017))); see also Intell. Ventures I LLC v. Capital One Fin. Corp., 850 F.3d 1332, 1342 (Fed. Cir. 2017) (similar); Elec. Power Grp., LLC v. Alstom S.A., 830 F.3d 1350, 1356 (Fed. Cir. 2016) (similar). Claiming a mere concept or functional result without disclosing the implementation details does not overcome USC 101. Applying an established technique to a new field or data set is insufficient for patent eligibility. Examiner further points out that improving efficiency by reducing computational time (pg. 15 of the applicant’s arguments of record and [0022] of specification) is not sufficient to show an improvement in computer functionality as set forth by the courts in Intellectual Ventures I LLC v. Capital One Bank (USA), 792 F.3d 1363, 1367, 115 USPQ2d 1636, 1639 (Fed. Cir. 2015); claiming the improved speed or efficiency inherent with applying the abstract idea on a computer does not integrate a judicial exception into a practical application or provide an inventive concept. This further supports the examiner’s assertion that the claims do not integrate a judicial exception into a practical application. To show an involvement of a computer assists in improving technology, the claims must recite details regarding how a computer aids the method, the extent to which the computer aids the method, or the significance of a computer to the performance of the method. Merely adding generic computer components to perform the method is not sufficient. Thus, the claim must include more than mere instructions to perform the method on a generic component or machinery to qualify as an improvement to an existing technology (MPEP 2106.05(a)(II)). In Finjan, Inc. v. Blue Coat Systems the courts found that the claims were “directed to a non-abstract improvement in computer functionality…” (MPEP 2106.04(d)). The present invention clearly does not meet the condition set forth by the courts and thus is not integrated into a practical application.
An analysis was performed under Step 2B, with court case citations, which didn’t result in the claim being eligible under USC 101. In comparison to Bascom, examiner points out that Bascom is not like the present application because Bascom claimed a technical improvement in the art i.e., a technology-based solution to filter content on the internet while the present application is not presenting an improvement to computer technology (as indicated above). There is nothing specialized about using off-the-shelf machine learning techniques on new data. The MPEP makes it clear that the search for a 101 inventive concept is distinct from demonstrating 102 novelty. In addition, the search for an inventive concept is different from an obviousness analysis under 35 U.S.C. 103. See, e.g., BASCOM Global Internet v. AT&T Mobility LLC, 827 F.3d 1341, 1350, 119 USPQ2d 1236, 1242 (Fed. Cir. 2016). Because they are separate and distinct requirements from eligibility, patentability of the claimed invention under 35 U.S.C. 102 and 103 with respect to the prior art is neither required for, nor a guarantee of, patent eligibility under 35 U.S.C. 101. The use of a computer or other machinery in its ordinary capacity for economic or other tasks or simply adding a general-purpose computer or computer components after the fact to an abstract idea does not provide significantly more. See Affinity Labs v. DirecTV, 838 F.3d 1253, 1262, 120 USPQ2d 1201, 1207 (Fed. Cir. 2016) (cellular telephone); TLI Communications LLC v. AV Auto, LLC, 823 F.3d 607, 613, 118 USPQ2d 1744, 1748 (Fed. Cir. 2016) (computer server and telephone unit). The applicant has not demonstrated that their invention is inventive. There is no justification to withdraw the USC 101. Therefore, the USC 101 rejection is strongly maintained
Regarding the USC 103 rejection, applicant’s arguments have been considered but are moot in light of newly cited reference: Nozaki.
Prior Art Cited but Not Relied Upon
Halimi, I., Piffo, E., Boudersa, O., Vilmorin, Y. M. C., Ait-ikhlef, M., Kone, K., ... & Dorcelus, G. (2026). A Latent Risk-Aware Machine Learning Approach for Predicting Operational Success in Clinical Trials based on TrialsBank. arXiv preprint arXiv:2603.29041.
This reference is relevant because it conceptually discloses the applicant’s invention.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WINSTON FURTADO whose telephone number is (571)272-5349. The examiner can normally be reached Monday-Friday 8:00 AM to 4:00 PM EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Mamon Obeid can be reached at (571) 270-1813. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/WINSTON R FURTADO/Primary Examiner, Art Unit 3687
