# US 2021/0241469 A1 · Unsupervised anchor handling for machine vision system > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2021/0241469 A1 - Document ID: US20210241469A1 - Status: Published application - Kind: A1 - Application number: 16/782,405 - Filed: 2020-02-05 - Published: 2021-08-05 - Earliest claimed: 2020-02-05 - Assignees: Datalogic IP Tech S.R.L. - Applicants: Datalogic I.P. Tech S.R.L. - Inventors: Martino Alessandrini - Web page: https://patentgrep.com/patent/US20210241469A1 ## CPC - G06V 10/25: Determination of region of interest [ROI] or a volume of interest [VOI] - G06T 7/33: using feature-based methods - G06V 30/19133: {Interactive pattern learning with a human teacher} - G06V 30/1914: {Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries, e.g. user dictionaries} - G06V 30/424: Postal images, e.g. labels or addresses on parcels or postal envelopes - G06T 2207/20081: Training; Learning - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract A device includes an image sensor and a processor to: receive training images that each include multiple visual features and an ROI; receive indications of locations of ROIs within each training image; perform at least one transform on the multiple visual features and ROI of at least one training image to align the multiple visual features and ROIs among all of the training images to identify a common set of visual features present within all of the training images; derive a converged ROI from at least a portion of the ROI of at least one training image; and generate an anchor model based on the converged ROI and the common set of visual features, wherein the common set of visual features defines the anchor and are each specified relative to the converged ROI, and the anchor model is used to derive a location of a candidate ROI in an image. ## Claims ### Claim 1 A device comprising: an image sensor configured to capture images; storage configured to store image data representing the captured images; and a processor communicatively coupled to at least the storage, the processor configured, during a training mode, to perform operations comprising: receive a set of training images captured by the image sensor, wherein each captured training image of the set of captured training images comprises multiple visual features and a region of interest (ROI); receive manually input indications of a location of the ROI within each captured training image of the set of captured training images; perform at least one of a rotational transform, a translational transform, a scaling transform or a shear transform on the multiple visual features and ROI of at least one captured training image of the set of captured training images to at least partially align the multiple visual features and ROIs among all of the captured training images of the set of captured training images to identify a common set of visual features that are present within all of the captured training images; derive a converged ROI based on at least a portion of the ROI of at least one of the captured training images of the set of captured training images; and generate an anchor model based on a combination of the converged ROI and the common set of visual features, wherein: the common set of visual features defines an anchor; a location of each visual feature is specified relative to the converged ROI; and the anchor model is to be used by the processor during an operating mode to derive a location of a candidate ROI relative to the anchor in a captured image. ### Claim 2 (depends on 1) The device of claim 1, wherein: the multiple visual features of each captured training image of the set of captured training images comprises multiple line segments; and the common set of visual features comprises the line segments of the multiple line segments within each of the captured training images that are present within all of the captured training images. ### Claim 3 (depends on 1) The device of claim 1, comprising an input device, wherein the processor is caused to operate at least the input device to provide a user interface (UI) by which the indications of a location of the ROI within each captured training image of the set of captured training images is manually provided. ### Claim 4 (depends on 1) The device of claim 1, wherein the performance of at least one of a rotational transform, a translational transform, a scaling transform or a shear transform comprises a performance, by the processor, of operations comprising: select a first captured training image of the set of captured training images to serve as a reference image; and perform the at least one of a rotational transform, a translational transform, a scaling transform or a shear transform on the multiple visual features and ROI of a second captured training image of the set of captured training images to maximize a number of visual features of the multiple of visual features present within the second captured training image that are aligned with corresponding visual features of the multiple visual features present within the first captured training image. ### Claim 5 (depends on 4) The device of claim 4, wherein the processor is further configured to cease the performance of the at least one of a rotational transform, a translational transform, a scaling transform or a shear transform in response to reaching a predetermined limit selected from a group consisting of: a minimum number of corresponding visual features aligned between the first captured training image and the second captured training image; and a maximum number of performances of at least one of a rotational transform, a translational transform, a scaling transform or a shear transform. ### Claim 6 (depends on 1) The device of claim 1, wherein the derivation of the converged ROI comprises a performance, by the processor, of operations comprising reshape an area overlapped by the ROIs of all of the captured training images of the set of captured training images to give the converged ROI a rectangular shape. ### Claim 7 (depends on 1) The device of claim 1, wherein use of the anchor model by the processor during the operating mode to derive a location of a candidate ROI in the captured image comprises a performance, by the processor, of operations comprising: search within the captured image for the common set of visual features of the anchor; and in response to the common set of visual features being found within the captured image, use an indication of the location of at least one visual feature of the common set of visual features relative to the converged ROI to derive the location of the candidate ROI within the captured image, wherein the processor is configured to perform operations comprising: perform at least one machine vision task within the candidate ROI; and transmit data output by the performance of the at least one machine vision task to another device via a network. ### Claim 8 A machine vision system comprising: a camera, wherein: the camera comprises an image sensor configured to capture images; and the camera is configured to use an anchor model, during an operating mode of the machine vision system, to derive a location of a candidate region of interest (ROI) relative to an anchor in a captured image; a manual input device; and a control device communicatively coupled to at least the camera and the input device, and comprising a processor configured, during a training mode of the machine vision system, to perform operations comprising: receive, from the camera, a set of training images captured by the image sensor, wherein each captured training image of the set of captured training images comprises multiple visual features and an ROI; receive, from the input device, manually input indications of a location of the ROI within each captured training image of the set of captured training images; perform at least one of a rotational transform, a translational transform, a scaling transform or a shear transform on the multiple visual features and ROI of at least one captured training image of the set of captured training images to at least partially align the multiple visual features and ROIs among all of the captured training images of the set of captured training images to identify a common set of visual features that are present within all of the captured training images; derive a converged ROI based on at least a portion of the ROI of at least one of the captured training images of the set of captured training images; and generate the anchor model based on a combination of the converged ROI and the common set of visual features, wherein: the common set of visual features defines the anchor; and a location of each visual feature is specified relative to the converged ROI. ### Claim 9 (depends on 8) The machine vision system of claim 8, wherein: the multiple visual features of each captured training image of the set of captured training images comprises multiple line segments; and the common set of visual features comprises the line segments of the multiple line segments within each of the captured training images that are present within all of the captured training images. ### Claim 10 (depends on 8) The machine vision system of claim 8, comprising a display, wherein the processor is caused to operate at least the input device and the display to provide a user interface (UI) to prompt an operator to operate at least the input device to manually provide the indications of a location of an ROI within each captured training image of the set of captured training images. ### Claim 11 (depends on 8) The machine vision system of claim 8, wherein the performance of at least one of a rotational transform, a translational transform, a scaling transform or a shear transform on at least one training image of the set of captured training images comprises a performance, by the processor, of operations comprising: select a first captured training image of the set of captured training images to serve as a reference image; and perform the at least one of a rotational transform, a translational transform, a scaling transform or a shear transform on the multiple visual features and ROI of a second captured training image of the set of captured training images to maximize a number of visual features of the multiple of visual features present within the second captured training image that are aligned with corresponding visual features of the multiple visual features present within the first captured training image. ### Claim 12 (depends on 11) The machine vision system of claim 11, wherein the processor is further configured to cease the performance of the at least one of a rotational transform, a translational transform, a scaling transform or a shear transform in response to reaching a predetermined limit selected from a group consisting of: a minimum number of corresponding visual features aligned between the first captured training image and the second captured training image; and a maximum number of performances of at least one of a rotational transform, a translational transform, a scaling transform or a shear transform. ### Claim 13 (depends on 8) The machine vision system of claim 8, wherein derivation of the converged ROI comprises a performance, by the processor, of operations comprising reshape an area overlapped by the ROIs of all of the captured training images of the set of captured training images to give the converged ROI a rectangular shape. ### Claim 14 (depends on 8) The machine vision system of claim 8, wherein use of the anchor model by the camera during the operating mode to derive a location of a candidate ROI in the captured image comprises a performance, by the camera, of operations comprising: search within the captured image for the common set of visual features of the anchor; and in response to the common set of visual features being found within the captured image, use an indication of the location of at least one visual feature of the common set of visual features relative to the converged ROI to derive the location of the candidate ROI within the captured image, wherein the camera is configured to perform operations comprising: perform at least one machine vision task within the candidate ROI; and transmit data output by the performance of the at least one machine vision task to another device via a network. ### Claim 15 A method comprising, during a training mode, performing operations comprising: receiving, at a processor and from an image sensor, a set of training images captured by the image sensor, wherein each captured training image of the set of captured training images comprises multiple visual features and a region of interest (ROI); receiving, at the processor, manually input indications of a location of the ROI within each captured training image of the set of captured training images; performing, by the processor, at least one of a rotational transform, a translational transform, a scaling transform or a shear transform on the multiple visual features and ROI of at least one captured training image of the set of captured training images to at least partially align the multiple visual features and ROIs among all of the captured training images of the set of captured training images to identify a common set of visual features that are present within all of the captured training images; deriving, by the processor, a converged ROI based on at least a portion of the ROI of at least one of the captured training images of the set of captured training images; and generating, by the processor, an anchor model based on a combination of the converged ROI and the common set of visual features, wherein: the common set of visual features defines an anchor; a location of each visual feature is specified relative to the converged ROI; and the anchor model is to be used by the processor during an operating mode to derive a location of a candidate ROI relative to the anchor in a captured image. ### Claim 16 (depends on 15) The method of claim 15, wherein: the multiple visual features of each captured training image of the set of captured training images comprises multiple line segments; and the common set of visual features comprises the line segments of the multiple line segments within each of the captured training images that are present within all of the captured training images. ### Claim 17 (depends on 15) The method of claim 15, wherein the performance of at least one of a rotational transform, a translational transform, a scaling transform or a shear transform comprises performing operations comprising: selecting, by the processor, a first captured training image of the set of captured training images to serve as a reference image; and performing, by the processor, the at least one of a rotational transform, a translational transform, a scaling transform or a shear transform on the multiple visual features and ROI of a second captured training image of the set of captured training images to maximize a number of visual features of the multiple of visual features present within the second captured training image that are aligned with corresponding visual features of the multiple visual features present within the first captured training image. ### Claim 18 (depends on 17) The method of claim 17, comprising ceasing, by the processor, the performance of the at least one of a rotational transform, a translational transform, a scaling transform or a shear transform in response to reaching a predetermined limit selected from a group consisting of: a minimum number of corresponding visual features aligned between the first captured training image and the second captured training image; and a maximum number of performances of at least one of a rotational transform, a translational transform, a scaling transform or a shear transform. ### Claim 19 (depends on 15) The method of claim 15, wherein the derivation of the converged ROI comprises performing operations comprising reshaping an area overlapped by the ROIs of all of the captured training images of the set of captured training images to give the converged ROI a rectangular shape. ### Claim 20 (depends on 15) The method of claim 15, wherein use of the anchor model by the processor during the operating mode to derive a location of a candidate ROI in the captured image comprises performing operations comprising: searching, by the processor, within the captured image for the common set of visual features of the anchor; and in response to the common set of visual features being found within the captured image, using, by the processor, an indication of the location of at least one visual feature of the common set of visual features relative to the converged ROI to derive the location of the candidate ROI within the captured image, wherein the processor is configured to perform operations comprising: performing, by the processor, at least one machine vision task within the candidate ROI; and transmitting, from the processor, data output by the performance of the at least one machine vision task to another device via a network.