# US 2022/0284694 A1 · Sampling for feature detection in image analysis > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2022/0284694 A1 - Document ID: US20220284694A1 - Status: Published application - Kind: A1 - Application number: 17/751,518 - Filed: 2022-05-23 - Published: 2022-09-08 - Earliest claimed: 2018-04-05 - Assignees: Imagination Technologies Limited - Applicants: Imagination Technologies Limited - Inventors: Timothy Smith - Web page: https://patentgrep.com/patent/US20220284694A1 ## CPC - G06V 30/2504: {Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches} - G06F 18/22: Matching criteria, e.g. proximity measures - G06V 10/426: Graphical representations - G06V 10/443: {by matching or filtering} - G06V 10/46: Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features (colour feature extraction G06V10/56) - G06V 10/462: {Salient features, e.g. scale invariant feature transforms [SIFT]} - G06V 10/52: Scale-space analysis, e.g. wavelet analysis (multi-scale boundary representations G06V10/42) - H04N 19/117: Filters, e.g. for pre-processing or post-processing (sub-band filter banks H04N19/635) - H04N 19/132: Sampling, masking or truncation of coding units, e.g. adaptive resampling, frame skipping, frame interpolation or high-frequency transform coefficient masking - G06T 2207/20016: Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform - G06V 10/467: {Encoded features or binary features, e.g. local binary patterns [LBP]} - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract A computer-implemented method for generating a feature descriptor for a location in an image for use in performing descriptor matching in analysing the image, the method comprising determining a set of samples characterising a location in an image by sampling scale-space data representative of the image, the scale-space data comprising data representative of the image at a plurality of length scales; and generating a feature descriptor in dependence on the determined set of samples. ## Claims ### Claim 1 A descriptor generation system for generating a feature descriptor for a location in an image for use in performing descriptor matching in analysing the image, the descriptor generation system being configured to: obtain scale-space data representative of the image, said scale-space data comprising a pre-filtered representation of the image at a plurality of length scales; the descriptor generation system having a feature descriptor generator configured to: determine a set of samples characterising a location in an image by sampling, using a sampling unit, the scale-space data representative of the image in accordance with a descriptor pattern of the feature descriptor, wherein the descriptor pattern is used to determine length scales at which the samples are to be sampled from the pre-filtered representation of the image; and generate the feature descriptor in dependence on the determined set of samples. ### Claim 2 (depends on 1) The descriptor generation system according to claim 1, in which the feature descriptor generator is further configured such that if a determined length scale does not match one of the length scales of the pre-filtered representation of the image then said sampling the scale-space data comprises interpolating between data at levels in the scale-space data associated with length scales above and below the determined length scale. ### Claim 3 (depends on 1) The descriptor generation system according to claim 1, in which the pre-filtered representation of the image has been filtered using one or more of a low-pass filter, a Gaussian filter and a box filter. ### Claim 4 (depends on 1) The descriptor generation system according to claim 1, in which the location in the image is one or more of: a point in the image; a pixel location in the image; and a keypoint in the image. ### Claim 5 (depends on 1) The descriptor generation system according to claim 1, in which the feature descriptor generator is configured to sample the scale-space data by: analysing portions of the scale-space data representing the location at different length scales to determine a measure of likelihood for each analysed portion, the measure of likelihood representing the likelihood of a feature being at the respective length scale; determining the portion of the scale-space data resulting in a turning point in the measure of likelihood; determining a length scale at which to sample the scale-space data to determine the set of samples in dependence on the determined portion. ### Claim 6 (depends on 1) The descriptor generation system according to claim 1, in which the descriptor generation system is configured to place a relatively smaller scale-space representation within a relatively larger scale-space representation in dependence on an identified length scale. ### Claim 7 (depends on 6) The descriptor generation system according to claim 6, in which the relatively larger scale-space representation comprises an image pyramid. ### Claim 8 (depends on 1) The descriptor generation system according to claim 1, in which the descriptor generation system is configured to identify the location in the image in accordance with one or more location identification or detection algorithms, and in which the descriptor generation system is configured to identify at least one of a location and a length scale in the scale-space data associated with the identified location in the image. ### Claim 9 (depends on 8) The descriptor generation system according to claim 8, in which the feature descriptor generator is configured to sample the scale-space data one or more of: in dependence on the identified length scale; at a level in the scale-space data associated with the identified length scale; and by interpolating between data at levels in the scale-space data associated with length scales above and below the identified length scale. ### Claim 10 (depends on 1) The descriptor generation system according to claim 1, in which the data in the scale-space data having been filtered at different length scales corresponds to filtered samples to be extracted in respect of different radial distances in the descriptor pattern from the centre of the descriptor pattern. ### Claim 11 (depends on 1) The descriptor generation system according to claim 1, in which the descriptor pattern comprises at least one ring surrounding the location in the image, in which the at least one ring is one of a circle, a wavy circle and a polygon. ### Claim 12 (depends on 1) The descriptor generation system according to claim 1, in which the descriptor generation system is configured to one or more of: store the determined set of samples in an array, and the feature descriptor generator is configured to generate the feature descriptor in dependence on the determined set of samples by forming a modified array; and determine a measure of rotation for the location in the image, the measure of rotation describing an angle between an orientation of the image and a characteristic direction of the image at the location, and generate the feature descriptor in dependence on the determined measure of rotation. ### Claim 13 (depends on 12) The descriptor generation system according to claim 12, in which the descriptor generation system is configured to form the modified array by one or more of: shifting elements of at least one portion of the array along a number of positions in the respective portion of the array, the number of positions being determined in dependence on the determined measure of rotation; and interpolating between two or more samples of the determined set of samples. ### Claim 14 (depends on 13) The descriptor generation system according to claim 13, in which the two or more samples of the determined set of samples have been obtained from one or more ring of the descriptor pattern. ### Claim 15 (depends on 14) The descriptor generation system according to claim 14, in which the descriptor generation system is configured to form the modified array by interpolating between the two or more samples of the determined set of samples along a portion of the shape of the ring to which the two or more samples correspond. ### Claim 16 (depends on 13) The descriptor generation system according to claim 13, in which the two or more samples of the determined set of samples between which interpolation is performed have been obtained from adjacent rings in the descriptor pattern. ### Claim 17 (depends on 13) The descriptor generation system according to claim 13, in which the two or more samples of the determined set of samples comprise N1 samples from a first ring and N2 samples from a second ring, where N1≤N2, in which the first ring is radially inside the second ring. ### Claim 18 (depends on 12) The descriptor generation system according to claim 12, in which the descriptor generation system is configured to discard the modified array once the feature descriptor has been generated. ### Claim 19 A computer-implemented method for generating a feature descriptor for a location in an image for use in performing descriptor matching in analysing the image, the method comprising: obtaining scale-space data representative of the image, said scale-space data comprising a pre-filtered representation of the image at a plurality of length scales; determining a set of samples characterising a location in an image by sampling the scale-space data representative of the image in accordance with a descriptor pattern of the feature descriptor, wherein the descriptor pattern is used to determine length scales at which the samples are to be sampled from the pre-filtered representation of the image; and generating the feature descriptor in dependence on the determined set of samples. ### Claim 20 A non-transitory computer readable storage medium having stored thereon computer readable instructions that, when executed at a computer system, cause the computer system to: obtain scale-space data representative of the image, said scale-space data comprising a pre-filtered representation of the image at a plurality of length scales; determine a set of samples characterising a location in an image by sampling the scale-space data representative of the image in accordance with a descriptor pattern of a feature descriptor, wherein the descriptor pattern is used to determine length scales at which the samples are to be sampled from the pre-filtered representation of the image; and generate the feature descriptor in dependence on the determined set of samples.