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 .
Response to Amendment
Claims 1, 3-6, 11, and 14-16 are amended. Claims 1-20 are pending.
Response to Arguments
Applicant's arguments filed 1/12/2026 have been fully considered.
Regarding the objections to claims 1, 4-6, 11, and 14-16, and as noted by Applicant on page 7 of the response, the amendments to claims 1, 4-6, 11, and 14-16 overcome the objections, which are withdrawn.
Regarding the rejections of claims 3, 6, and 16 under 112(b), and as explained by Applicant on pages 7-8 of the response, the amendments to claims 3, 6, and 16 overcome the rejections, which are withdrawn.
Regarding the rejections of claims 1-20 under 101, the Examiner respectfully disagrees with Applicant’s arguments that the amendments to independent claims 1 and 11 overcome the rejections for the following reasons.
On pages 8-9 of the response, Applicant contends that the feature “wherein the final linear regression model is validated against a validation data set comprising ground truth data not utilized to calculate the initial linear regression model,” entails a technological improvement to the field of computer vision function. Applicant cites the December 5, 2025 USPTO Memorandum regarding 101 guidelines and in particular reference by the Memorandum to Ex Parte Desjardins, Appeal No. 2024-000567 (Memo). Applicant notes on page 8 that the Memo indicates that improvements to computer functionality versus being directed to an abstract idea are an important factor in 101 eligibility determinations, and further notes on page 9 that MPEP 2106.04(d), subsection III is amended to reference Desjardins as an example of eligible subject matter in part due to “benefits including reduced storage, reduced system complexity and streamlining, and preservation of performance attributes associated with earlier tasks during subsequent computational tasks as technological improvements that were discussed in the patent application specification.” Applicant further notes the Memo indicates that improvement in how machine learning functions “in operation” may be eligible and that further indicates that Examiners should not evaluate claims at an unduly high level of generality that results in dismissal of potentially meaningful technical limitations.
Regarding application of the guidelines set forth in the Memo to the Applicant’s claims 1 and 11, Applicant contends on page 9 of the response that claims 1 and 11 are directed to the type of technological improvements that the Memo and Desjardins cite as eligible under 101, namely improvements in how the claimed computer vision systems and methods (forms of AI) function. In support, Applicant cites Applicant’s specification as disclosing that the selection of validation data as exclusive of data used for generating the model (training data) results in improved ability of the computer vision system/method to accurately and reliably determine roof ages.
The Examiner submits that “wherein the final linear regression model is validated against a validation data set comprising ground truth data not utilized to calculate the initial linear regression model” itself falls within the mathematical concepts judicial exception. As explained in Applicant’s specification in paragraph [0026], validation (apparently entailed in either or both of steps 40 and 42) entails statistical (inherently mathematical) functions. Furthermore, the improvement Applicant refers to is an improvement in accuracy of results, not an improvement in terms of how a computer, and a computer-implemented computer vision system, actually operates/functions. More specifically, using a validation data set that is in some way optimized (e.g., exclusive of data that has already been used for model training so that that validation function is more robust in terms of determining whether the training was sufficient to result in a model that accurately predicts/estimates roof ages) does not in any way improve the manner in which the computer system functions. Therefore, the Examiner finds that the feature added by amendment itself falls within the judicial exception and does not integrate the other elements falling within the judicial exception into a practical application such that the rejections of claims 1 and 11 and all claims depending therefrom are maintained.
Regarding the rejections of claims 1-20 under 103, the Examiner respectfully disagrees with Applicant’s arguments on page 10 of the response that the amendments overcome the rejections for the following reasons.
On page 10 of the response, contends that the prior arts including Christopulous teach or suggest validating a final linear regression model against a validation set comprising ground truth data not utilized to calculate an initial linear regression model as recited in claims 1 and 11. The Examiner submits that Christopulous discloses this element in paragraph [0033] (verification module 320 may use the same data as used in training (classification features used for training as described in [0032]), or alternatively, verification module 320 may verify the training results (validate model) using “other data with known characteristics” (i.e., other ground truth data)).
Therefore, the rejections of claims 1 and 11 under 103 unpatentable over Christopulos (US 2014/0270492 A1) in view of Grant (US 9,152,863 A1), and in further view of Lovings (US 2024/0303795 A1) are maintained.
Claim Objections
Claims 1 and 11 are objected to because of the following informalities:
In each of claims 1 and 11, the conjunction “and” should be inserted following the next-to-last claim element ending with “to calculate the initial linear regression model.”
Appropriate correction is required.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention in each of these claims is directed to the abstract idea judicial exception without significantly more.
Representative independent claim 1 recites:
“[a] computer vision system for determining an age or a remaining life of a roof of a structure, comprising:
a processor in communication with at least one data source and an end-user device, the processor programmed to:
receive first and second images from the at least one data source depicting an area of interest of the roof;
process the first and second images to detect at least one change in a condition of the roof over time;
calculate a ground truth model of a roof age and a roof condition based on the at least one change;
calculate an initial linear regression model using the ground truth model;
filter noise from the initial linear regression model;
calculate and validate a final linear regression model from the filtered initial linear regression model, wherein the final linear regression model is validated against a validation set comprising ground truth data not utilized to calculate the initial linear regression model; and
determine an age or a remaining life of the roof of the structure using the final linear regression model.”
The claim limitations considered to fall within in the abstract idea are highlighted in bold font above and the remaining features are “additional elements.”
Step 1 of the subject matter eligibility analysis entails determining whether the claimed subject matter falls within one of the four statutory categories of patentable subject matter identified by 35 U.S.C. 101: process, machine, manufacture, or composition of matter. Claim 1 recites a system and independent claim 11 recites a method and each therefore falls within a statutory category.
Step 2A, Prong One of the analysis entails determining whether the claim recites a judicial exception such as an abstract idea. Under a broadest reasonable interpretation, the highlighted portions of claim 1 fall within the abstract idea judicial exception. Specifically, under the 2019 Revised Patent Subject matter Eligibility Guidance, the highlighted subject matter falls within the mental processes category (including an observation, evaluation, judgment, opinion) and the mathematical concepts category (mathematical relationships, mathematical formulas or equations, mathematical calculations). MPEP § 2106.04(a)(2).
The recited functions:
“receive first and second images from the at least one data source depicting an area of interest of the roof;
process the first and second images to detect at least one change in a condition of the roof over time;
calculate a ground truth model of a roof age and a roof condition based on the at least one change,”
“filter noise from the initial linear regression model;” and
“determine an age or a remaining life of the roof of the structure,”
may be performed as mental processes.
Receiving first and second images from at least one data source depicting an area of interest of the roof may be performed via mental processes (e.g., observation such as via computer display). Processing the first and second images to detect at least one change in a condition of the roof over time may be performed via mental processes (e.g., evaluation of the images and judgment to determine changes in roof condition). Calculating a ground truth model of a roof age and a roof condition based on the at least one change may be performed via mental processes (e.g., evaluation of the determined changes and judgment to determine ground truth data associating the changes with other data such as age data associated with the images). Filtering noise from the initial linear regression model may be performed via mental processes (i.e., evaluating the data in a ground truth dataset/model to selectively identify outliers for removal prior to being input to linear regression model). Determining an age or a remaining life of a roof of a structure may be performed via mental processes (e.g., judgment).
The type of high-level information analysis and deduction recited in these elements has been found by the Federal Circuit to constitute patent ineligible matter (see Electric Power Group v. Alstom, S.A., 830 F.3d 1350, 1353-54, 119 USPQ2d 1739, 1741-42 (Fed. Cir. 2016), a claim to "collecting information, analyzing it, and displaying certain results of the collection and analysis," where the data analysis steps are recited at a high level of generality such that they could practically be performed in the human mind).
The recited functions:
“calculate an initial linear regression model using the ground truth model;
filter noise from the initial linear regression model;
calculate and validate a final linear regression model from the filtered initial linear regression model, wherein the final linear regression model is validated against a validation set comprising ground truth data not utilized to calculate the initial linear regression model;” and
determine an age or a remaining life of the roof of the structure “using the final linear regression model,”
are determined by the Examiner as falling within the mathematical relationships sub-category of mathematical concepts (MPEP 2106.04(a)(2)) because a linear regression model is fundamentally characterized by mathematical relationships (models mathematically linear relationship characterizing the correlation between variables). Therefore, the steps of calculating a linear regression model, filtering noise from the model (i.e., adjusting what constitutes the model), calculating and validating the model, and finally using the model to determine an age or remaining life all entail mathematical relationships. Regarding “wherein the final linear regression model is validated against a validation set comprising ground truth data not utilized to calculate the initial linear regression model,” Applicant’s specification in paragraph [0026] describes validation (apparently entailed in either or both of steps 40 and 42) as entailing statistical (inherently mathematical) functions. The Examiner further notes that the type of information, in terms of being part of the overall set of ground truth data that was not used for calculating the initial regression model, does change the character of this step in terms of falling within the mathematical concepts exception. The Examiner further notes that even if this element is interpreted to functionally entail a form of selection (e.g., selecting a subset of overall ground truth data for use in validation versus being used for training), such selection would fall within the mental processes judicial exception (e.g., judgement).
Step 2A, Prong Two of the analysis entails determining whether the claim includes additional elements that integrate the recited judicial exception into a practical application. “A claim that integrates a judicial exception into a practical application will apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that the claim is more than a drafting effort designed to monopolize the judicial exception” (MPEP § 2106.04(d)).
MPEP § 2106.04(d) sets forth considerations to be applied in Step 2A, Prong Two for determining whether or not a claim integrates a judicial exception into a practical application. Based on the individual and collective limitations of claim 1 and applying a broadest reasonable interpretation, the most applicable of such considerations appear to include: improvements to the functioning of a computer, or to any other technology or technical field (MPEP 2106.05(a)); applying the judicial exception with, or by use of, a particular machine (MPEP 2106.05(b)); and effecting a transformation or reduction of a particular article to a different state or thing (MPEP 2106.05(c)).
Regarding improvements to the functioning of a computer or other technology, none of the “additional elements” including “a computer vision system” comprising “a processor in communication with at least one data source and an end-user device” in any combination appear to integrate the abstract idea in a manner that technologically improves any aspect of a device or system that may be used to implement the highlighted steps or a device for implementing the highlighted steps such as a signal processing device or a generic computer. Instead, “a computer vision system” (a broadest reasonable interpretation of which entails a computer system that processes image information in some interpretive manner) comprising “a processor in communication with at least one data source and an end-user device” constitutes a high-level characterization of a conventional computer configuration for processing image data that inherently require connectivity to image data sources and typically include I/O interfaces for connectivity with end user devices.
Regarding application of the judicial exception with, or by use of, a particular machine, the additional element “a computer vision system” comprising “a processor in communication with at least one data source and an end-user device” is configured and implemented in a conventional rather than a particularized manner of implementing roof age monitoring.
Regarding a transformation or reduction of a particular article to a different state or thing, claim 1 does not include any such transformation or reduction. Instead, claim 1 as a whole entails receiving input information (first and second images), applying standard processing techniques (computer processor) to the information to determine condition change information for a roof and to construct and refine (filter) a mathematical construct (linear regression model) and to apply the mathematical construct to determine age or remaining life of a roof with the additional elements failing to provide a meaningful integration of the abstract idea (mental processes and mathematical relationships) in an application that transforms an article to a different state. Instead, the additional elements represent extra-solution activity that do not integrate the judicial exception into a practical application.
Regarding “receive first and second images from the at least one data source depicting an area of interest of the roof,” even if this element is interpreted to fall outside the mental processes exception (observation), it constitutes high-level data gathering implemented via conventional computer processing operations (input/output for receiving input information) and therefore constitutes extra-solution activity that fails to integrate the judicial exception into a practical application.
Regarding “calculate a ground truth model of a roof age and a roof condition based on the at least one change,” even if this element is interpreted to fall outside the mental processes exception, it constitutes routine and high-level data collection (collecting data related to the determination of a change in condition of the roof over time to form the dataset/model) implemented via conventional computer processing operations and therefore constitutes extra-solution activity that fails to integrate the judicial exception into a practical application.
In view of the various considerations encompassed by the Step 2A, Prong Two analysis, claim 1 does not include additional elements that integrate the recited abstract idea into a practical application.
Therefore, claim 1 is directed to a judicial exception and requires further analysis under Step 2B.
Regarding Step 2B, and as explained in the Step 2A Prong Two analysis, the additional elements in claim 1 constitute extra solution activity and therefore do not result in the claim as a whole amounting to significantly more than the judicial exception. Furthermore, the additional elements appear to be generic and well understood as evidenced by the disclosures of Christopulos (US 2014/0270492 A1) and Grant (US 9,152,863), each of which teach virtually the same computer-based image processing (computer vision) structure.
As explained in the grounds for rejecting claim 1 under 103, Christopulos teaches “a computer vision system” comprising “a processor in communication with at least one data source and an end-user device,” as does Grant (FIG. 1 computer system 200 configured to receive and process image data from satellites 120; FIG. 2 computer system 200 includes processor 205 communicatively connected to a network interface 215 and user interface 220 and includes roof condition assessment module 254 within memory 210; FIG. 5 image processing including blocks 405, 425, 430, and Has Roof Been Repaired, Replaced, or Damaged? (interpretive processing of image data)).
Therefore, the additional elements are insufficient to amount to significantly more than the judicial exception.
Claim 1 is therefore not patent eligible.
Independent claim 11 includes substantially the same features constituting a judicial exception as claim 1 and does not recite additional elements that either integrate the judicial exception into a practical application or result in the claim as a whole amounting to significantly more than the judicial exception.
Claim 11 is therefore also not patent eligible for the same reasons as set forth for claim 1.
Dependent claims 2-10 and 12-20 provide additional features/steps which are part of an expanded algorithm that includes the abstract idea of claims 1 and 11 (Step 2A, Prong One). None of dependent claims 2-10 and 12-20 recite additional elements that integrate the abstract idea into practical application (Step 2A, Prong Two), and all fail the “significantly more” test under the step 2B for substantially similar reasons as discussed with regards to the independent claims.
Claim 2, representatively also of claim 12, recites the additional element “wherein the processor is programmed to display the age or the remaining life of the roof on a graphical interface screen superimposed over an aerial image of the roof or the area of interest,” which entails conventional computer display processing that constitutes extra-solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claims 3 and 13 are found to fall within the mental processes exception because identifying an area of interest using a boundary such as zip code, county name, climate division, or state may be performed via mental processes (e.g., evaluation and judgment). The use of a processor for implementing the function, as indicated in the grounds for rejecting claim 1 constitutes extra solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claims 4 and 14 recite an extension of the step of “process the first and second images to detect at least one change in a condition of the roof over time” and are therefore found to fall within the mental processes exception.
Claim 5, representative also of claim 15, recites the additional element “wherein the processor detects the at least one change using computer vision neural network,” which entails conventional computer vision processing that as explained with regard to claim 1 constitutes extra-solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception. Use of a neural network further constitutes extra-solution activity because it merely represents a known computer implemented programming type to effectuate the judicial exception.
Claims 6, representatively also of claim 16, recites the additional element “wherein the ground truth model is determined at least in part by aggregating a plurality of roof ages each having an associated confidence level,” which entails high-level and conventional computer-based data collection (aggregating data). Therefore, the additional elements recited by claims 6 and 16 constitute extra-solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claims 7 and 17 recite an extension of the step of “filter noise from the initial linear regression model” and that does not include additional elements that integrate the judicial exception into a practical application or result in the claim as a whole amounting to significantly more than the judicial exception, such that each of claims 7 and 17 are found to fall within the mental processes exception and mathematical relationships exception.
Claims 8 and 18 recites an application of the step of using the final linear regression model with no additional elements and therefore each fall within the mathematical relationships sub-category of the mathematical concepts exception.
Claims 9 and 19 are found to fall within the mental processes exception because expressing a remaining life of a roof in terms of risk corresponding to a range of years fundamentally constitutes a determination of the remaining roof life in such terms, which may be performed via mental processes (e.g., evaluation of data indicating remaining life and judgment in determining a characterization in terms of risk corresponding to a range of years). Even if the elements of claims 9 and 19 are interpreted to fall outside the mental processes exception, these elements constitute routine computer output processing in terms of output result characterization and therefore constitutes extra-solution activity neither integrates the judicial exception into a practical application nor result in the claim as a whole amounting to significantly more than the judicial exception.
Claims 10 and 20 are found to fall within the mental processes exception because calibrating/adjusting modeling of the age or remaining life of a roof using at least one of updated imagery, updated roof age data, or user feedback may be performed via mental processes (e.g., evaluation of updated imagery data to determine a calibration/adjustment of the age or remaining life data).
Dependent claims 2-10 and 12-20 therefore also constitute ineligible subject matter.
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, 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 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Christopulos (US 2014/0270492 A1) in view of Grant (US 9,152,863 A1), and in further view of Lovings (US 2024/0303795 A1).
As to claim 1, Christopulos teaches “[a] computer vision system (FIG. 1 computer-implemented system 101 for performing image analysis as set for in FIG. 5) for determining an age or a remaining life of a roof of a structure ([0014] system determines age of building or property, [0016] age of roof structure to be assessed, [0030] data age-labeled (i.e., for model that will correspondingly classify); [0039] structure, which may be roof, classified in accordance with age), comprising:
a processor (FIG. 1 CPU 102 and GPU 104 within system 101) in communication with at least one data source (FIG. 1 CPU 102 and GPU 104 part of communicatively integrated system 101 that includes image sensor 121 (provides image data), I/O unit 103 (provides input information from external sources), NIM 108, and data storage 116) and an end-user device (FIG. 1 CPU 102 and GPU 104 part of communicatively integrated system 101 that includes I/O unit 103 and display 118; [0020]-[0021], [0025]), the processor programmed to:
receive first and second images ([0023] and [0025] images received; [0025]-[0026] image data includes current image data (first image) with previously stored image data (second image) to determine physical differences between older and newer roof tiles) from the at least one data source ([0023] images assessed may be received from storage and/or per [0025] from I/O unit 103. Image information originates and is therefore received from image sensor 121) depicting an area of interest of” [a] “roof ([0025] images describe structure to be assessed (i.e., of interest) that per [0016] may be a roof (roof itself of interest) in which the image contains areas of particular interest such as shingles and that per [0026] may correspond in terms of currently imaged roof tiles compared with previously imaged stock tiles of a roof);
process the first and second images to detect at least one change in a condition of the roof over time ([0016] and [0025]-[0026] image data processed via filtering and feature extraction including comparing current image data with previously stored images to determine physical differences between older and newer roof tiles);
calculate a ground truth model of a roof age and a roof condition based on the at least one change ([0027]-[0028] the determinations of physical differences utilized to label data (labeled data constituting a ground truth model) to reflect which physical characteristics in association with various states of aging);
calculate an initial” “regression model using the ground truth model (FIG. 3 training module 117; [0030] training module 117 trains analysis system 101 based on extracted features that per [0027]-[0028] are labelled (ground truth) to assess images to generate algorithm module 310; [0032] algorithm module 310 may be configured as any number of classifier models including logistic regression);
filter noise from the initial” “regression model (FIG. 2 filtering module 210 within extraction module 115; [0025] filtering module 210 filters out irrelevant and/or unexpected data points (each of irrelevant and unexpected data undesirable and therefore noise) from the data that is processed by extraction unit 115 that forms the labeled dataset (dataset used for training) that is encompassed within the scope of the overall model (as interpreted in view of Applicant’s specification and claim 7) that ultimately results in training/generation of algorithm module 310);
calculate and validate a final” “regression model from the filtered initial linear regression model ([0025] filtering out irrelevant and/or unexpected data points from the data that is processed by extraction unit 115 results in a determined/calculated (final) labeled dataset that ultimately results in training/generation of algorithm module 310 in which the model is validated at least in part via the removal of the irrelevant and/or unexpected data points. In a distinct aspect of validation, FIG. 3 depicts a verification module 320 configured to verify (validate) training results for the model, [0033]), wherein the final linear regression model is validated against a validation data set comprising ground truth data not utilized to calculate the initial linear regression model ([0033] verification module 320 may use the same data as used in training (classification features used for training as described in [0032]), or alternatively, verification module 320 may verify the training results (validate model) using “other data with known characteristics” (i.e., other ground truth data)); and
determine an age or a remaining life of the roof of the structure using the final” “regression model ([0014] system that per [0030] is trained using training module 117 determines age of building or property, [0016] age of roof structure to be assessed, [0030] data age-labeled (i.e., for model that will correspondingly classify); [0039] structure, which may be roof, classified in accordance with age).
As explained above, Christopolus teaches processing first and second images of a structure/building and further teaches such as in [0025]-[0026] that a first image may be may correspond in terms of currently imaged roof tiles and a second image may correspond with previously imaged stock tiles of a corresponding type of structure (e.g., a corresponding type of roof). However, Christopulos does not appear to expressly teach that the first and second images that are processed to detect a change in condition over time are obtained from the same structure/building.
Grant discloses a system for assessing roof condition in which first and second images of the same roof are processed to determine a change in condition of the roof over time (Abstract damage assessed based on first and second images of roof at different times; col. 2 lines 5-14 first and second images of roof at different times used to determine whether roof repaired or replaced; col. 8 lines 8-21 describing collection of images at different times for same roof).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Grant’s teaching of using first and second images from a same roof to determine changes to the roof over time to the system taught by Christopolus, which teaches that comparable structures are utilized to assess time-related changes, such that in combination the system is configured to obtain the first and second images from the same roof.
The motivation would have been to optimize the accuracy of comparison between like objects (e.g., roof type, tiles, etc.) to consequently improve the accuracy of the subsequent damage assessment related processing as suggested by each of Grant and Christopolus.
Christopulos teaches that any one of a variety of different data models including a logistical regression model may be generated/trained and utilized for determining roof age ([0030] and [0032]). However, Christopulos does not specifically list a linear regression model as one of the potential model types, such that Christopulos does not expressly teach calculating and filtering noise from an initial “linear regression model,” and calculating and validate a final “linear regression model” to be used for determine an age of the roof.
Lovings discloses a system for implementing a trained model for analyzing structural damage that occurs over time to a roof using historical image data that includes roof age (Abstract; [0029]-[0031]; [0058]-[0059]) in which the modelling for such age related change in condition may be performed using a linear regression model ([0046] photographic data may be processed by machine learning model to determine a structure status such as age; [0120] modeling may be implemented as a linear regression model).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Lovings teaching of training and utilizing a linear regression model for analyzing structural damage to a roof to the system taught by Christopulos, which teaches that any of a variety of different types of learning models including regression-type models may be calculated and utilized for determining age based on condition, such that the combined system is configured to calculate and filter noise from an initial linear regression model, and calculating and validate a final linear regression model to be used for determine an age of the roof.
Such a combination would amount to applying a known design option for modeling historical image data indicating age-related changes to a roof condition to achieve predictable results.
As to claim 2, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1, wherein the processor is programmed to display the age or the remaining life of the roof (Christopulos: [0047] assessment module 112 may output as a display assessment results that per [0039] may include damage and/or age of the structure) on a graphical interface screen superimposed over an” “image of the roof or the area of interest (Christopulos: [0047] the assessment results may be overlaid (superimposed) on an image of the structure that has been assessed. Examiner notes graphical interface screen (e.g., display screen/object that provides output information) is inherent to a computer display).”
Christopulos appears largely silent regarding the means/mechanism for originally obtaining the images and therefore does not expressly teach that the roof image is an “aerial” image.
Grant further teaches that the images obtained for roof assessment may be aerial images (FIG. 1 satellite 120 configured to obtain aerial images of roof 110; col. 5 lines 4-23; col. 6 lines 46-51).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Grant’s teaching of using aerial imagery to obtain the roof images to the system taught by Christopulos as modified by Grant and Lovings such that in combination the roof image is an aerial image over which the graphical interface screen is superimposed.
Such a combination would amount to applying a known design option for obtaining roof images to achieve predictable results.
As to claim 3, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1,” and Christopulos further teaches “wherein the processing is programmed to identify an area of interest using a boundary ([0018] assessment area may be a property or geographic area (area inherently bounded)).”
Grant further teaches that the assessment location may designated in terms of zip code (col. 6 lines 58-62; col. 9 line 63 through col. 10 line 3).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Grant’s teaching of using zip code as an imagery assessment location designator to the system taught by Christopulos in which an assessment area bounding is determined in accordance with a geographic area such that in combination the processor is programmed to identify an area of interest using a boundary (bounded geographic area) including a zip code.
The motivation would have been to associate a geographic area to be assessed in terms of information in the form of a zip code that is useful in terms of associating the collected imagery data to particular properties that are typically associated with zip codes as suggested by Grant.
As to claim 4, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1, wherein the at least one change in the condition of the roof includes one or more of roof discoloration (Christopulos: [0026] change in condition may be difference in color such as due to weather aging), percentage of missing material, structural damage percentage, percentage of the roof covered by a tarp, percentage of roof debris, percentage of the roof that is anomalous, or percentage of patched or repaired roof sections.”
As to claim 5, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1,” and Chrisopulos further teaches a computer vision neural network for determining changes to a roof ([0030] neural network is among the modeling options for classifying images in terms of age, damage condition, color, etc. Examiner notes that neural network configured to classify images constitutes a broadest reasonable interpretation of computer vision neural network in view of Applicant’s specification).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Christopulos’s teaching that a computer vision neural network may be used for determining a change in condition (e.g., condition indicating damage such as color) to Christopulos generalized teaching of determining at least one change for purposes of generating a ground truth models ([0016] and [0025]-[0026] describing generalized comparison of images to determine a change in condition) such that a computer vision neural network model is utilized for the initial ground truth model generation step such that “the processor detects the at least one change using a computer vision neural network.”
Such a combination would amount to applying a known design option for determining an age-related change in condition of a roof to achieve predictable results.
As to claim 6, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1, wherein the ground truth model is determined at least in part by aggregating a plurality of roof ages each having an associated confidence level (Christopulos: [0015]-[0016] data points that may include age are labelled (i.e., data points prepared as training data and therefore having an associated level of confidence); [0030]).”
As to claim 7, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1, wherein the noise is filtered from the initial regression model by removing outliers from the ground truth model (Christopulos: FIG. 2 filtering module 210 within extraction module 115; [0025] filtering module 210 filters out (removes from further analysis) irrelevant and/or unexpected (outlier) data points (filtering per [0025] is performed prior to processing by reference module 220 and feature module, which assemble the final labeled/training dataset (ground truth model) as indicated in [0026]-[0027]) that are above or below a selected confidence level (Christopulos: [0025] the filtering may be performed by comparing with pre-defined data (entails a selected reference/confidence level) describing the structure).”
As to claim 8, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1, wherein the final linear regression model models at least one stage of roof deterioration including initial slow deterioration, intermediate accelerated deterioration, and final decelerated deterioration (Christopulos model as modified by Lovings to be configured as a linear regression model. Per Christopulos FIG. 5 blocks 504 and 505, [0045] assessment system receives second data representative of structure that is classified (per [0039] and [0044] classifications may be damage and/or age (stage of deterioration). Examiner notes that such classification is based on image data inherently taken at some time that would inevitably coincide with a particular stage of deterioration, including deterioration that happens to be initial slow, intermediate accelerated, or final decelerated).”
As to claim 9, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1,” and the Examiner notes that per claim 1, “wherein the remaining life of the roof is expressed as a level of risk corresponding to a range of years of life remaining for the roof” is an optional element (per claim 1 the system determines an age or a remaining life of the roof).
However, for the purpose of advancing prosecution, Grant further teaches determining a remaining life of a roof wherein the remaining life of the roof is expressed as a level of risk corresponding to a range of years of life remaining (col. 13 lines 30-34, determine remaining useful life of roof based on age in which remaining useful life is expressed as 10 years and/or 50%. Examiner notes that 10 years entails a risk level associated with a range year 1 to year 10 from the present).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Grant’s teaching of determining a remaining life of a roof wherein the remaining life of the roof is expressed as a level of risk corresponding to a range of years of life remaining to the system taught by Christopulos as modified by Grant and Lovings such that a remaining life determination is expressed as a level of risk corresponding to a range of years of life remaining for the roof.
The motivation would have been to provide remaining life information that is useful for downstream processing or user consideration that usefully conveys a risk over time to the roof based on the remaining life data as suggested by Grant.
As to claim 10, the combination of Christopulos, Grant, and Lovings teaches “[t]he system of Claim 1, wherein the processor is programmed to calibrate modeling of the age of the roof or the remaining life of the roof using at least one of updated imagery (Christopulos: [0030] initial and followup training may be performed by algorithm module 310. Examiner notes that followup training using images constitutes use of followup (updated) imagery (i.e., post-initial training imagery)), updated roof age data, or user feedback.”
As to claim 11, Christopulos teaches “[t] computer vision method (FIG. 1 method implemented by computer-implemented system 101 for performing image analysis as set for in FIG. 5) for determining an age or a remaining life of a roof of a structure ([0014] system determines age of building or property, [0016] age of roof structure to be assessed, [0030] data age-labeled (i.e., for model that will correspondingly classify); [0039] structure, which may be roof, classified in accordance with age), comprising:
receiving at a processor (FIG. 1 CPU 102 and GPU 104 within system 101) first and second images ([0023] and [0025] images received; [0025]-[0026] image data includes current image data (first image) with previously stored image data (second image) to determine physical differences between older and newer roof tiles) from at least one data source (FIG. 1 CPU 102 and GPU 104 part of communicatively integrated system 101 that includes image sensor 121 (provides image data), I/O unit 103 (provides input information from external sources), NIM 108, and data storage 116; [0023] images assessed may be received from storage and/or per [0025] from I/O unit 103. Image information originates and is therefore received from image sensor 121) depicting an area of interest of” [a] “roof ([0025] images describe structure to be assessed (i.e., of interest) that per [0016] may be a roof (roof itself of interest) in which the image contains areas of particular interest such as shingles and that per [0026] may correspond in terms of currently imaged roof tiles compared with previously imaged stock tiles of a roof);
processing the first and second images to detect at least one change in a condition of the roof over time ([0016] and [0025]-[0026] image data processed via filtering and feature extraction including comparing current image data with previously stored images to determine physical differences between older and newer roof tiles);
calculating a ground truth model of a roof age and a roof condition based on the at least one change ([0027]-[0028] the determinations of physical differences utilized to label data (labeled data constituting a ground truth model) to reflect which physical characteristics in association with various states of aging);
calculating an initial” “regression model using the ground truth model (FIG. 3 training module 117; [0030] training module 117 trains analysis system 101 based on extracted features that per [0027]-[0028] are labelled (ground truth) to assess images to generate algorithm module 310; [0032] algorithm module 310 may be configured as any number of classifier models including logistic regression);
filtering noise from the initial” “regression model (FIG. 2 filtering module 210 within extraction module 115; [0025] filtering module 210 filters out irrelevant and/or unexpected data points (each of irrelevant and unexpected data undesirable and therefore noise) from the data that is processed by extraction unit 115 that forms the labeled dataset (dataset used for training) that is encompassed within the scope of the overall model (as interpreted in view of Applicant’s specification and claim 7) that ultimately results in training/generation of algorithm module 310);
calculating and validating a final” “regression model from the filtered initial linear regression model ([0025] filtering out irrelevant and/or unexpected data points from the data that is processed by extraction unit 115 results in a determined/calculated (final) labeled dataset that ultimately results in training/generation of algorithm module 310 in which the model is validated at least in part via the removal of the irrelevant and/or unexpected data points. In a distinct aspect of validation, FIG. 3 depicts a verification module 320 configured to verify (validate) training results for the model, [0033]), wherein the final linear regression model is validated against a validation data set comprising ground truth data not utilized to calculate the initial linear regression model ([0033] verification module 320 may use the same data as used in training (classification features used for training as described in [0032]), or alternatively, verification module 320 may verify the training results (validate model) using “other data with known characteristics” (i.e., other ground truth data));
determining an age or a remaining life of the roof of the structure using the final” “regression model ([0014] system that per [0030] is trained using training module 117 determines age of building or property, [0016] age of roof structure to be assessed, [0030] data age-labeled (i.e., for model that will correspondingly classify); [0039] structure, which may be roof, classified in accordance with age).
As explained above, Christopolus teaches processing first and second images of a structure/building and further teaches such as in [0025]-[0026] that a first image may be may correspond in terms of currently imaged roof tiles and a second image may correspond with previously imaged stock tiles of a corresponding type of structure (e.g., a corresponding type of roof). However, Christopulos does not appear to expressly teach that the first and second images that are processed to detect a change in condition over time are obtained from the same structure/building.
Grant discloses a system/method for assessing roof condition in which first and second images of the same roof are processed to determine a change in condition of the roof over time (Abstract damage assessed based on first and second images of roof at different times; col. 2 lines 5-14 first and second images of roof at different times used to determine whether roof repaired or replaced; col. 8 lines 8-21 describing collection of images at different times for same roof).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Grant’s teaching of using first and second images from a same roof to determine changes to the roof over time to the method taught by Christopolus, which teaches that comparable structures are utilized to assess time-related changes, such that in combination the system is configured to obtain the first and second images from the same roof.
The motivation would have been to optimize the accuracy of comparison between like objects (e.g., roof type, tiles, etc.) to consequently improve the accuracy of the subsequent damage assessment related processing as suggested by each of Grant and Christopolus.
Christopulos teaches that any one of a variety of different data models including a logistical regression model may be generated/trained and utilized for determining roof age ([0030] and [0032]). However, Christopulos does not specifically list a linear regression model as one of the potential model types, such that Christopulos does not expressly teach calculating and filtering noise from an initial “linear regression model,” and calculating and validate a final “linear regression model” to be used for determine an age of the roof.
Lovings discloses a system/method for implementing a trained model for analyzing structural damage that occurs over time to a roof using historical image data that includes roof age (Abstract; [0029]-[0031]; [0058]-[0059]) in which the modelling for such age related change in condition may be performed using a linear regression model ([0046] photographic data may be processed by machine learning model to determine a structure status such as age; [0120] modeling may be implemented as a linear regression model).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Lovings teaching of training and utilizing a linear regression model for analyzing structural damage to a roof to the method taught by Christopulos, which teaches that any of a variety of different types of learning models including regression-type models may be calculated and utilized for determining age based on condition, such that the combined method is configured to calculate and filter noise from an initial linear regression model, and calculating and validate a final linear regression model to be used for determine an age of the roof.
Such a combination would amount to applying a known design option for modeling historical image data indicating age-related changes to a roof condition to achieve predictable results.
As to claim 12, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11, further comprising displaying the age or the remaining life of the roof (Christopulos: [0047] assessment module 112 may output as a display assessment results that per [0039] may include damage and/or age of the structure) on a graphical interface screen superimposed over an” “image of the roof or the area of interest (Christopulos: [0047] the assessment results may be overlaid (superimposed) on an image of the structure that has been assessed. Examiner notes graphical interface screen (e.g., display screen/object that provides output information) is inherent to a computer display).”
Christopulos appears largely silent regarding the means/mechanism for originally obtaining the images and therefore does not expressly teach that the roof image is an “aerial” image.
Grant further teaches that the images obtained for roof assessment may be aerial images (FIG. 1 satellite 120 configured to obtain aerial images of roof 110; col. 5 lines 4-23; col. 6 lines 46-51).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Grant’s teaching of using aerial imagery to obtain the roof images to the method taught by Christopulos as modified by Grant and Lovings such that in combination the roof image is an aerial image over which the graphical interface screen is superimposed.
Such a combination would amount to applying a known design option for obtaining roof images to achieve predictable results.
As to claim 13, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11,” and Christopulos further teaches “identifying an area of interest using a boundary ([0018] assessment area may be a property or geographic area (area inherently bounded)).”
Grant further teaches that the assessment location may designated in terms of zip code (col. 6 lines 58-62; col. 9 line 63 through col. 10 line 3).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Grant’s teaching of using zip code as an imagery assessment location designator to the method taught by Christopulos in which an assessment area bounding is determined in accordance with a geographic area such that in combination the area of interest is identified using a boundary (bounded geographic area) including a zip code.
The motivation would have been to associate a geographic area to be assessed in terms of information in the form of a zip code that is useful in terms of associating the collected imagery data to particular properties that are typically associated with zip codes as suggested by Grant.
As to claim 14, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11, wherein the at least one change in the condition of the roof includes one or more of roof discoloration (Christopulos: [0026] change in condition may be difference in color such as due to weather aging), percentage of missing material, structural damage percentage, percentage of the roof covered by a tarp, percentage of roof debris, percentage of the roof that is anomalous, or percentage of patched or repaired roof sections.”
As to claim 15, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11,” and Chrisopulos further teaches a computer vision neural network for determining changes to a roof ([0030] neural network is among the modeling options for classifying images in terms of age, damage condition, color, etc. Examiner notes that neural network configured to classify images constitutes a broadest reasonable interpretation of computer vision neural network in view of Applicant’s specification).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Christopulos’s teaching that a computer vision neural network may be used for determining a change in condition (e.g., condition indicating damage such as color) to Christopulos generalized teaching of determining at least one change for purposes of generating a ground truth models ([0016] and [0025]-[0026] describing generalized comparison of images to determine a change in condition) such that a computer vision neural network model is utilized for the initial ground truth model generation step such that “the processor detects the at least one change using a computer vision neural network.”
Such a combination would amount to applying a known design option for determining an age-related change in condition of a roof to achieve predictable results.
As to claim 16, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11, wherein the ground truth model is determined at least in part by aggregating a plurality of roof ages each having an associated confidence level (Christopulos: [0015]-[0016] data points that may include age are labelled (i.e., data points prepared as training data and therefore having an associated level of confidence); [0030]).”
As to claim 17, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11, wherein the noise is filtered from the initial regression model by removing outliers from the ground truth model (Christopulos: FIG. 2 filtering module 210 within extraction module 115; [0025] filtering module 210 filters out (removes from further analysis) irrelevant and/or unexpected (outlier) data points (filtering per [0025] is performed prior to processing by reference module 220 and feature module, which assemble the final labeled/training dataset (ground truth model) as indicated in [0026]-[0027]) that are above or below a selected confidence level (Christopulos: [0025] the filtering may be performed by comparing with pre-defined data (entails a selected reference/confidence level) describing the structure).”
As to claim 18, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11, wherein the final linear regression model models at least one stage of roof deterioration including initial slow deterioration, intermediate accelerated deterioration, and final decelerated deterioration (Christopulos model as modified by Lovings to be configured as a linear regression model. Per Christopulos FIG. 5 blocks 504 and 505, [0045] assessment system receives second data representative of structure that is classified (per [0039] and [0044] classifications may be damage and/or age (stage of deterioration). Examiner notes that such classification is based on image data inherently taken at some time that would inevitably coincide with a particular stage of deterioration, including deterioration that happens to be initial slow, intermediate accelerated, or final decelerated).”
As to claim 19, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11,” and the Examiner notes that per claim 11, “wherein the remaining life of the roof is expressed as a level of risk corresponding to a range of years of life remaining for the roof” is an optional element (per claim 11 the method determines an age or a remaining life of the roof).
However, for the purpose of advancing prosecution, Grant further teaches determining a remaining life of a roof wherein the remaining life of the roof is expressed as a level of risk corresponding to a range of years of life remaining (col. 13 lines 30-34, determine remaining useful life of roof based on age in which remaining useful life is expressed as 10 years and/or 50%. Examiner notes that 10 years entails a risk level associated with a range year 1 to year 10 from the present).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Grant’s teaching of determining a remaining life of a roof wherein the remaining life of the roof is expressed as a level of risk corresponding to a range of years of life remaining to the method taught by Christopulos as modified by Grant and Lovings such that a remaining life determination is expressed as a level of risk corresponding to a range of years of life remaining for the roof.
The motivation would have been to provide remaining life information that is useful for downstream processing or user consideration that usefully conveys a risk over time to the roof based on the remaining life data as suggested by Grant.
As to claim 20, the combination of Christopulos, Grant, and Lovings teaches “[t]he method of Claim 11, further comprising calibrating modeling of the age of the roof or the remaining life of the roof using at least one of updated imagery (Christopulos: [0030] initial and followup training may be performed by algorithm module 310. Examiner notes that followup training using images constitutes use of followup (updated) imagery (i.e., post-initial training imagery)), updated roof age data, or user feedback.”
Conclusion
THIS ACTION IS MADE FINAL. 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 MATTHEW W BACA whose telephone number is (571)272-2507. The examiner can normally be reached Monday - Friday 8:00 am - 5:30 pm.
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, Andrew Schechter can be reached at (571) 272-2302. 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.
/MATTHEW W. BACA/Examiner, Art Unit 2857
/ANDREW SCHECHTER/Supervisory Patent Examiner, Art Unit 2857