DETAILED ACTION
Response to Arguments
Applicant’s amendment filed on May 15, 2025 is acknowledged. Currently Claims 1, 4-6, 8, 11-13, 15 and 18-28 are pending. Claims 21-28 are new.
Applicant's arguments with respect independent claims 1, 8 and 15 have been considered but are moot in view of the new ground(s) of rejection. Amended claims 1, 8 and 15 results in a different scope than that of the originally presented Claims 1, 8 and 15 respectively.
Claim Rejections - 35 USC § 103
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 of this title, 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.
Claims 1, 8, 15, 21, 24, and 27 are rejected under 35 U.S.C. 103 as being unpatentable over Son et al US 2022/0172352 hereinafter referred to as Song in view of Ramirez et al. US2019/0228527 hereinafter referred to as Ramirez and Ulmann et al. US2022/0004737 hereinafter referred to as Ulmann.
As per Claim 1, Son teaches a method of determining a density of cells in a cell image, the method comprising: acquiring a cell image; (Son, Paragraph [0055], “The user interface 130 may include an input interface configured to collect various cell images applied to the drug evaluation apparatus 100 so as to evaluate a drug and configured to input a user request and commands for evaluating the drug therethrough. The cell images may be input by the user or may be obtained from the server 300”)
extracting mapped features of the cell image by an autoencoder, wherein the mapped features (Son, Paragraph[0060]-[0061], has disclosed using a trained learning model to process the input image data (such as “TensorFlow JS”), wherein the autoencoder “feature extraction” process is performed by the CNN architecture corresponding to a GAP process and convolutional layers Son, Paragraph [0073]-[0075], wherein the features are extracted for mapping to a plurality of classes/categories corresponding to at least the density ranges (Son, Paragraph [0061 “configured to set the number of the concentrations of a drug to be evaluated, a drug concentration inputter 142 configured to input the respective concentrations of the drug depending on the set number of the concentrations of the drug to be evaluated”).
inputting the mapped features into a neural network classifier and obtaining a feature category (Son, Paragraph [0061], [0074]-[0075], A concentration type category is the classification determined by the system “molar concentration” of the input cell images based on the learned CNN type neural network)
outputting the density range; (Son, Paragraph [0068]-[0069], The density of the cell images “molar concentration” is output by the neural network process)
Son does not explicitly teach wherein the mapped features describe feature distribution information of the cell image in different feature categories,
wherein the feature category comprises a unique identification identifying one category;
Ramirez teaches wherein the mapped features describe feature distribution information of the cell image in different feature categories, (Ramirez, Figure 2, Paragraph [0022], “In another aspect, provided is a method of determining a classification of a particle in a biological sample by mapping a subset of extracted features into a cascade classifier architecture, the mapping including using a first level machine learning model to compare the subset of extracted features to a previously stored data set, wherein the extracted features may be extracted from images, calculating a probability value using the first level machine learning model, comparing the probability value to a predetermined comparison table, determining a cell classification if the probability value is at or above a threshold value, using a second level machine learning model if the probability value is below the threshold value, creating an ascending sorted list of values according to their classification performance in relation to a cell or particle category using the second level machine learning model, combining the probability value and the sorted list [0023] of values to create a second level score, using the second level score to determine a cell classification.”)
wherein the feature category comprises a unique identification identifying one category; (Ramirez, Figure 2, Paragraph [0022])
Thus it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to implement the teachings of Ramirez into Son because by using extracted features with use of machine learning model will allow for different and accurate categorization/classification of the biological sample.
Son in view of Ramirez does not explicitly teach obtaining the density range corresponding to the feature category;
Ulmann teaches teach obtaining the density range corresponding to the feature category; (Ulmann, Paragraph [0041], “According to an embodiment, the classifier may consult a database of known image coefficients associated with known cancer cell types and known genetic signatures. For example, an image coefficient such as a specific vacuole density, size, or count, or a range thereof, may be associated with breast cancer, prostate cancer, or any other type of cancer, and/or may be associated with a sub-type of breast cancer, prostate cancer, or any other type of cancer. Accordingly, when a cell is imaged and the vacuole density, size, or count falls within the specific count or the range, the classifier identifies the cell as being a specific type or sub-type of cancer cell”)
Thus it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to implement the teachings of Ulmann into Son in view of Ramirez because obtaining the density ranges of different categories of cells will provide additional information in regards to the classified cells determined by Son in view of Ramirez.
Therefore it would have been obvious to one of ordinary skill to combine the three references to obtain the invention in Claim 1.
As per Claim 8, Claim 8 claims an electronic device performing the method as claimed in Claim 1. Therefore the rejection and rationale are analogous to that made in Claim 1.
As per Claim 15, Claim 15 claims a non-transitory storage medium having stored thereon at least one computer-readable instructions that, when the at least one computer-readable instructions are executed by a processor to implement a method as claimed in Claim 1. Therefore the rejection and rationale are analogous to that made in Claim 1.
As per Claim 21, Son in view of Ramirez and Ulmann teaches the method according to claim 1, wherein a process of training the autoencoder comprises: acquiring a plurality of sample images; (Son, Paragraph [0070]-[0071]) has disclosed training the learning model (comprising at least an “autoencoder” that is a GAP process of Son, Paragraph [0073]-[0075])inputting the plurality of sample images into a preset neural network; and(Son, Paragraph [0070]-[0072] has disclosed receiving by a pre-trained learning model which is then further refined using the training sample data comprising the input data and labels) training the preset neural network and obtaining the autoencoder. (Son, Paragraph [0071]-[0074], [0061], has disclosed training the layers of the neural network based on at least , a corresponding number of classes, features, number of classifications, and the input image data has disclosed inputting and training a neural network model based on a plurality of input image data corresponding to a plurality of groups/classes of predetermined density/concentration levels.)
The rationale applied to the rejection of claim 1 has been incorporated herein.
As per Claim 24, Claim 24 claims the same limitation as Claim 21 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 21.
As per Claim 27, Claim 27 claims the same limitation as Claim 21 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 21.
Claims 4-6, 11-13, 18-20, 22-23, 25-26, and 28 are rejected under 35 U.S.C. 103 as being unpatentable over Son et al US 2022/0172352 hereinafter referred to as Song in view of Ramirez et al. US2019/0228527 hereinafter referred to as Ramirez and Ulmann et al. US2022/0004737 hereinafter referred to as Ulmann as applied to Claims 21, 24 and 27 respectively and further in view of Deng CN 11898525A.
As per Claim 22, Song in view of Ramirez and Ulmann teaches the method according to claim 21,
Song in view of Ramirez and Ulmann does not explicitly teach wherein the plurality of sample images comprises a plurality of groups of the sample images, and densities of cells of the sample images in the same group belong to the same density range, and densities of cells of the sample images in different groups belong to different density ranges. (Deng, abstract, Paragraph [0098], [0101], [0110-0111] has disclosed a process of training a learning model for processing sets of image data to determine densities within input image data, wherein the training process groups each of the input training image data based upon the density/concentration range (Paragraph [0095]-[0097] – grouped according to density/concentration ranges)..)
Thus it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to implement the teachings of Deng into Song in view of Ramirez and Ulmann because Deng and Son are analogous art of image data processing to train a neural network to extract features from captured image data and output the concentration/density measurement based on a learned neural network model. The output of each prior art reference of Deng and Son corresponding to a number of trained classes of density/concentration levels. It would have been obvious to one of ordinary skill in the art at the time of the invention to at least try to modify the learning model receiving images classified into a set number of classes based on density taught by Son wherein the classes are not specified has having a range or a grouping process of respective ranges, with the process of density range identification and training wherein the samples are sorted into sets of corresponding density/concentration ranges as disclosed by Deng. The resulting combination being an organized set of labeled training images having a finite set of classes as required by the teachings of Son, when the input density data has an infinite set of possible values (inherent to a dataset over a set of real and non-discrete numbers), hence the infinite set of Son must be limited to a finite set of classes, thus a range of values of density must correspond to a class of Son, else there would be an infinite set of classes of Son corresponding to the infinite set of measured density values, and not the finite set of classes disclosed by Son (at least para 0061 and 0070-0071). The corresponding teachings of Deng disclosing how to limit a set of real-world continuous data set values of density/concentration to a finite number of groups such as a finite set of categories or classes of a neural network for classifying density data of input images as disclosed by each of the teachings of Son and Deng. The finite set of data being a range of values for each of a set finite number of categories. Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to modify the teachings of Son by grouping and setting the density values to a range of corresponding values such that the neural network processing and output correspond to a required finite set of density value classifications as required by each of the teachings of Deng and Son, wherein each disclose categorizing a continuous set of values into a finite number of categories/classes. Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to at least try to combine the teachings of Son and Deng to achieve the limitations of the present claim. Therefore, as a result of Son in view Deng: It has been taught that the processing of the continuous set of density/concentration of values as disclosed by each of Son and Deng is performed on a range of said values as disclosed by Deng, hence the processing of density/concentration of Son is with respect to ranges of density values corresponding to a finite set of classes.
Therefore it would have been obvious to one of ordinary skill to combine the four references to obtain the invention in Claim 22.
As per Claim 23, Song in view of Ramirez and Ulmann teaches the method according to claim 21,
Song in view of Ramirez and Ulmann does not explicitly teach wherein the mapped features of the sample images with similar density of cells are distributed with less variation, the mapped features of the sample images with different density of cells are distributed with greater variation.
Deng teaches wherein the mapped features of the sample images with similar density of cells are distributed with less variation, the mapped features of the sample images with different density of cells are distributed with greater variation. (Deng, abstract, Paragraph [0098], [0101], [0110-0111] disclosed a process of training a learning model for processing sets of image data to determine densities within input image data, wherein the training process groups each of the input training image data based upon the density/concentration range (Deng, Paragraph [0095]-[0097] – grouped according to density/concentration ranges)..)The input training samples are grouped with less variation for similar density because those of a similar density range are grouped into a same training sample grouping than those of different density of cell having greater variation. The greater variation corresponding to density values in other groups of said training sample frames corresponding to other density ranges.)
Thus it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to implement the teachings of Deng into Song in view of Ramirez and Ulmann because Deng and Son are analogous art of image data processing to train a neural network to extract features from captured image data and output the concentration/density measurement based on a learned neural network model. The output of each prior art reference of Deng and Son corresponding to a number of trained classes of density/concentration levels. It would have been obvious to one of ordinary skill in the art at the time of the invention to at least try to modify the learning model receiving images classified into a set number of classes based on density taught by Son wherein the classes are not specified has having a range or a grouping process of respective ranges, with the process of density range identification and training wherein the samples are sorted into sets of corresponding density/concentration ranges as disclosed by Deng. The resulting combination being an organized set of labeled training images having a finite set of classes as required by the teachings of Son, when the input density data has an infinite set of possible values (inherent to a dataset over a set of real and non-discrete numbers), hence the infinite set of Son must be limited to a finite set of classes, thus a range of values of density must correspond to a class of Son, else there would be an infinite set of classes of Son corresponding to the infinite set of measured density values, and not the finite set of classes disclosed by Son (at least para 0061 and 0070-0071). The corresponding teachings of Deng disclosing how to limit a set of real-world continuous data set values of density/concentration to a finite number of groups such as a finite set of categories or classes of a neural network for classifying density data of input images as disclosed by each of the teachings of Son and Deng. The finite set of data being a range of values for each of a set finite number of categories. Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to modify the teachings of Son by grouping and setting the density values to a range of corresponding values such that the neural network processing and output correspond to a required finite set of density value classifications as required by each of the teachings of Deng and Son, wherein each disclose categorizing a continuous set of values into a finite number of categories/classes. Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to at least try to combine the teachings of Son and Deng to achieve the limitations of the present claim. Therefore, as a result of Son in view Deng: It has been taught that the processing of the continuous set of density/concentration of values as disclosed by each of Son and Deng is performed on a range of said values as disclosed by Deng, hence the processing of density/concentration of Son is with respect to ranges of density values corresponding to a finite set of classes.
Therefore it would have been obvious to one of ordinary skill to combine the four references to obtain the invention in Claim 23.
As per Claim 4, Son in view Ramirez, Ulmann and Deng teaches the method according to claim 22, a process of training the neural network comprising:
inputting the plurality of groups of the sample images into the autoencoder to obtain mapped features corresponding to each group of the sample images; (Son, Paragraph [0073]-[0075], disclose mapping the features of the input image data to a set of classes of the output vector, wherein Deng, of Son in view Deng, has disclosed that the sets of input training images correspond to groups of classes of ranges of densities/concentrations.)
determining features distribution of all mapped features in different density ranges according to the mapped features corresponding to each group of the sample images and the density ranges corresponding to the plurality of groups of the sample images; (Son, Paragraph [0073]-[0075], has disclosed processing the features to assign the set of features to a finite sets of classes, wherein Deng further discloses that the classes correspond to range distributions of density/concentration values.)
obtaining an initial classifier; and (Son, Paragraph [0075]-[0076], the output of the initial SoftMax is processed by a MSE optimization process to update the parameters of the neural network, thus training and optimizing the neural network based on at least the training data and features contained therein.)
applying the features distribution to train the initial classifier and obtaining the neural network classifier. (Son, Paragraph [0075]-[0076] the output of the initial SoftMax is processed by a MSE optimization process to update the parameters of the neural network, thus training and optimizing the neural network based on at least the training data and features contained therein Paragraph [0078]-[0080] – optimization and update process of Son).)
The rationale applied to the rejection of claim 4 has been incorporated herein.
As per Claim 5, Son in view Ramirez, Ulmann and Deng teaches the method according to claim 4, wherein the neural network classifier comprises a fully connected layer and a SoftMax layer. (Son, Paragraph [0073]-[0075], The autoencoder of the neural network having a GAP or fully connected convolutional layer following by a SoftMax layer for assigning classes based upon the set of feature data of the GAP or fully connected convolutional type layers.)
The rationale applied to the rejection of claim 4 has been incorporated herein.
As per Claim 6, Son in view Ramirez, Ulmann and Deng teaches the method according to claim 5, wherein the fully connected layers calculates values of the type to which it belongs, according to the mapped features of the cell image, the SoftMax layer outputs the feature category. (Son, Paragraph [0073]-[0075], discloses a fully connected feature extraction process for generating class information for the received and extracted feature data, the reception of the data by a SoftMax layer, and the output of a corresponding class based on density and feature information. The SoftMax layer receiving a vector of values corresponding in size to the number of classes (density ranges – as per Son in view of Deng), wherein the values of said vector are inherently sets of probability values that the features extracted correspond to in said vector (i.e. the vector data structure of a process such as disclosed by Son is a 1-Dimensional vector of values wherein a class at a position in the vector is assigned the value at the respective position in the vector). Then the corresponding SoftMax process of Son assigns the class based on the set of probability values inherent to the process of learning disclosed by Son.)
The rationale applied to the rejection of claim 5 has been incorporated herein.
As per Claim 25, Claim 25 claims the same limitation as Claim 22 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 22.
As per Claim 26, Claim 26 claims the same limitation as Claim 23 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 23.
As per Claim 11, Claim 11 claims the same limitation as Claim 4 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 4.
As per Claim 12, Claim 12 claims the same limitation as Claim 5 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 5.
As per Claim 13, Claim 13 claims the same limitation as Claim 6 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 6.
As per Claim 28, Claim 28 claims the same limitation as Claim 22 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 22.
As per Claim 18, Claim 18 claims the same limitation as Claim 4 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 4.
As per Claim 19, Claim 19 claims the same limitation as Claim 5 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 5.
As per Claim 20, Claim 20 claims the same limitation as Claim 6 and is dependent on a similarly rejected independent claim. Therefore the rejection and rationale are analogous to that made in Claim 6.
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 MING HON whose telephone number is (571)270-5245. The examiner can normally be reached M-F 9am - 5pm.
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, Emily Terrell can be reached on 571-270-3717. 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.
/MING Y HON/Primary Examiner, Art Unit 2666