# US 11,113,572 B2 · Systems, methods, and storage media for evaluating images > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,113,572 B2 - Document ID: US11113572B2 - Status: Granted - Kind: B2 - Application number: 16/672,227 - Filed: 2019-11-01 - Published: 2020-08-13 - Granted: 2021-09-07 - Earliest claimed: 2019-02-08 - Assignees: Vizit Labs, Inc. - Applicants: Adhark, Inc. - Inventors: Jehan Hamedi; Zachary Halloran; Elham Saraee - Web page: https://patentgrep.com/patent/US11113572B2 ## CPC - G06V 10/761: {Proximity, similarity or dissimilarity measures} - G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting - G06F 18/22: Matching criteria, e.g. proximity measures - G06V 10/82: using neural networks - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Embodiments may: select a set of training images; extract a first set of features from each training image of the set of training images to generate a first feature tensor for each training image; extract a second set of features from each training image to generate a second feature tensor for each training image; reduce a dimensionality of each first feature tensor to generate a first modified feature tensor for each training image; reduce a dimensionality of each second feature tensor to generate a second modified feature tensor for each training image; construct a first generative model representing the first set of features and a second generative model representing the second set of features of the set of training images; identify a first candidate image; and apply a regression algorithm to the first candidate image and each of the first generative model and the second generative model to determine whether the first candidate image is similar to the set of training images. ## Claims ### Claim 1 A system comprising: one or more hardware processors including machine-readable instructions to: identify a first image; extract a first set of features from the first image to generate a first feature tensor for the first image; extract a second set of features from the first image to generate a second feature tensor for the first image; identify a second image; extract a third set of features from the second image to generate a third feature tensor for the second image; extract a fourth set of features from the second image to generate a fourth feature tensor for the second image; apply a first regression analysis to determine a first geometrical distance between the first feature tensor of the first image and the third feature tensor of the second image; apply a second regression analysis to determine a second geometrical distance between the second feature tensor of the first image and the fourth feature tensor of the second image; and determine a similarity between the first image and the second image based on the first geometrical distance and the second geometrical distance. ### Claim 2 (depends on 1) The system of claim 1, wherein the one or more hardware processors further include machine-readable instructions to calculate a similarity score representing a degree of visual similarity between the first image and the second image. ### Claim 3 (depends on 2) The system of claim 2, wherein the one or more hardware processors further include machine-readable instructions to calculate a uniqueness score of the first image with respect to the second image. ### Claim 4 (depends on 3) The system of claim 3, wherein the one or more hardware processors further include machine-readable instructions to calculate the uniqueness score of the first image by: calculating an inverse of the similarity score; and identifying the inverse as the uniqueness score. ### Claim 5 (depends on 1) The system of claim 1, wherein the one or more hardware processors further include machine-readable instructions to reduce a dimensionality of the first feature tensor prior to applying the first regression analysis to determine the first geometrical distance between the first feature tensor of the first image and the third feature tensor of the second image. ### Claim 6 (depends on 5) The system of claim 5, wherein the one or more hardware processors further include machine-readable instructions to apply a weight to the first set of features extracted from the first image to generate a set of weighted first features, wherein the first feature tensor is generated based on the set of weighted first features. ### Claim 7 (depends on 1) The system of claim 1, wherein: the first set of features extracted from the first image comprises object features; and the one or more hardware processors further include machine-readable instructions to extract the first set of features from the first image by: propagating data corresponding to the first image through at least one network including at least one of an object detection neural network, an object classification neural network, or an object recognition neural network, wherein the network comprises an input layer, a plurality of intermediate layers, and an output layer; and extracting outputs from at least one of the plurality of intermediate layers of the network. ### Claim 8 (depends on 1) The system of claim 1, wherein the one or more hardware processors further include machine-readable instructions to extract the first set of features from the first image by extracting at least one of a set of object features, a set of scene features, a set of intensity features, a set of contrast features, a set of color features, and a set of blurriness features from the first image. ### Claim 9 A method comprising: identifying a first image; extracting a first set of features from the first image to generate a first feature tensor for the first image; extracting a second set of features from the first image to generate a second feature tensor for the first image; identifying a second image; extracting a third set of features from the second image to generate a third feature tensor for the second image; extracting a fourth set of features from the second image to generate a fourth feature tensor for the second image; applying a first regression analysis to determine a first geometrical distance between the first feature tensor of the first image and the third feature tensor of the second image; applying a second regression analysis to determine a second geometrical distance between the second feature tensor of the first image and the fourth feature tensor of the second image; and determining a similarity between the first image and the second image based on the first geometrical distance and the second geometrical distance. ### Claim 10 (depends on 9) The method of claim 9, further comprising calculating a similarity score representing a degree of visual similarity between the first image and the second image. ### Claim 11 (depends on 10) The method of claim 10, further comprising calculating a uniqueness score of the first image with respect to the second image. ### Claim 12 (depends on 11) The method of claim 11, wherein calculating the uniqueness score of the first image comprises: calculating an inverse of the similarity score; and identifying the inverse as the uniqueness score. ### Claim 13 (depends on 9) The method of claim 9, further comprising reducing a dimensionality of the first feature tensor prior to applying the first regression analysis to determine the first geometrical distance between the first feature tensor of the first image and the third feature tensor of the second image. ### Claim 14 (depends on 13) The method of claim 13, further comprising applying a weight to the first set of features extracted from the first image to generate a set of weighted first features, wherein the first feature tensor is generated based on the set of weighted first features. ### Claim 15 (depends on 9) The method of claim 9, wherein the first set of features extracted from the first image comprises object features, and extracting the first set of features from the first image comprises: propagating data corresponding to the first image through at least one network including at least one of an object detection neural network, an object classification neural network, or an object recognition neural network, wherein the network comprises an input layer, a plurality of intermediate layers, and an output layer; and extracting outputs from at least one of the plurality of intermediate layers of the network. ### Claim 16 (depends on 9) The method of claim 9, wherein extracting the first set of features from the first image comprises extracting at least one of a set of object features, a set of scene features, a set of intensity features, a set of contrast features, a set of color features, and a set of blurriness features from the first image. ### Claim 17 A non-transitory computer-readable storage medium having instructions embodied thereon, the instructions being executable by one or more processors to perform a method comprising: identifying a first image; extracting a first set of features from the first image to generate a first feature tensor for the first image; extracting a second set of features from the first image to generate a second feature tensor for the first image; identifying a second image; extracting a third set of features from the second image to generate a third feature tensor for the second image; extracting a fourth set of features from the second image to generate a fourth feature tensor for the second image; applying a first regression analysis to determine a first geometrical distance between the first feature tensor of the first image and the third feature tensor of the second image; applying a second regression analysis to determine a second geometrical distance between the second feature tensor of the first image and the fourth feature tensor of the second image; and determining a similarity between the first image and the second image based on the first geometrical distance and the second geometrical distance. ### Claim 18 (depends on 17) The computer-readable storage medium of claim 17, wherein the method further comprises: calculating a similarity score representing a degree of visual similarity between the first image and the second image; calculating an inverse of the similarity score; and identifying the inverse as a uniqueness score of the first image with respect to the second image. ### Claim 19 (depends on 17) The computer-readable storage medium of claim 17, wherein the method further comprises: reducing a dimensionality of the first feature tensor prior to applying the first regression analysis to determine the first geometrical distance between the first feature tensor of the first image and the third feature tensor of the second image; and applying a weight to the first set of features extracted from the first image to generate a set of weighted first features, wherein the first feature tensor is generated based on the set of weighted first features. ### Claim 20 (depends on 17) The computer-readable storage medium of claim 17, wherein the first set of features extracted from the first image comprises object features, and extracting the first set of features from the first image comprises: propagating data corresponding to the first image through at least one network including at least one of an object detection neural network, an object classification neural network, or an object recognition neural network, wherein the network comprises an input layer, a plurality of intermediate layers, and an output layer; and extracting outputs from at least one of the plurality of intermediate layers of the network.