# US 12,182,196 B2 · Image retrieval system, image retrieval method, and storage medium > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,182,196 B2 - Document ID: US12182196B2 - Status: Granted - Kind: B2 - Application number: 17/640,687 - Filed: 2019-12-25 - Published: 2022-10-13 - Granted: 2024-12-31 - Earliest claimed: 2019-12-25 - Assignees: NEC Corporation - Applicants: NEC Corporation - Inventors: Toru Takahashi; Rui Ishiyama; Kengo Makino - Web page: https://patentgrep.com/patent/US12182196B2 ## CPC - G06F 16/532: Query formulation, e.g. graphical querying - G06F 16/583: using metadata automatically derived from the content - G06T 7/001: {using an image reference approach} - G06V 10/22: by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition - G06T 2207/20021: Dividing image into blocks, subimages or windows - G06T 2207/30108: Industrial image inspection - G06T 2207/30121: CRT, LCD or plasma display - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract This image retrieval system comprises: a storage means for storing a plurality of registered images capturing a plurality of objects to be registered; an acquisition means for acquiring a retrieval image including a region of interest; an extraction means for extracting a feature amount from a region, in the retrieval image, other than the region of interest; a retrieval means for retrieving, from the plurality of registered images stored in the storage means, a registered image having a feature amount that matches the feature amount extracted by the extraction means; and an output means for outputting, in a manner that makes it possible to recognize a region corresponding to the region of interest in the retrieval image, all or a portion of the registered image retrieved by the retrieval means. ## Claims ### Claim 1 An image retrieval system comprising: at least one memory storing instructions; and at least one processor configured to access the at least one memory and execute the instructions to: store, in a storage, a plurality of registered images obtained by capturing a plurality of objects to be registered; acquire a retrieval image including a region of interest; extract a feature from a region other than the region of interest of the retrieval image; retrieve, from the plurality of registered images stored in the storage, a registered image having a feature that matches the feature; and output some or all of the registered images in a mode that a region corresponding to the region of interest in the retrieval image is recognizable. ### Claim 2 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: output the some or all of the registered images, by superimposing an image in which the region corresponding to the region of interest of the retrieval image is recognizable. ### Claim 3 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: output the some or all of the registered images retrieved by the retrieval means, by superimposing an image in which a region corresponding to the region from which the feature of the retrieval image has been extracted is recognizable. ### Claim 4 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: output the registered image retrieved by the retrieval means and an enlarged image of the region in the registered image, the region corresponding to the region of interest in the retrieval image. ### Claim 5 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: output the some or all of the registered images including the region corresponding to the region of interest of the retrieval image, and the retrieval image. ### Claim 6 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: store information associated with each of the plurality of registered images in the storage, and output the information associated with the registered image. ### Claim 7 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: store, in the storage, registered images captured at a plurality of different timings in a distribution process of the objects to be registered, retrieve, from the plurality of registered images stored in the storage, a registered image at each of the plurality of timings, the registered image having a feature that matches the feature extracted by the extraction means, and output some or all of the registered images at the plurality of timings in a mode that a region corresponding to the region of interest in the retrieval image is recognizable. ### Claim 8 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: specify the region of interest of the retrieval image using the retrieval image. ### Claim 9 (depends on 8) The image retrieval system according to claim 8, wherein the at least one processor is further configured to execute the instructions to: divide the retrieval image into partial regions; and specify the region of interest by determining presence or absence of a defect in each of the divided partial regions. ### Claim 10 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: acquire information designating the region of interest; and specify the region of interest of the retrieval image based on the information designating the region of interest. ### Claim 11 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: extract the feature from a region other than the region of interest, in a predetermined region including the region of interest. ### Claim 12 (depends on 1) The image retrieval system according to claim 1, wherein the at least one processor is further configured to execute the instructions to: extract the feature from a predetermined region adjacent to the region of interest. ### Claim 13 An image retrieval method comprising: acquiring a retrieval image including a region of interest; extracting a feature from a region other than the region of interest of the retrieval image; retrieving, from a plurality of registered images obtained by capturing a plurality of objects to be registered, a registered image having a feature that matches the extracted feature; and outputting some or all of the retrieved registered images in a mode that a region corresponding to the region of interest in the retrieval image is recognizable. ### Claim 14 (depends on 13) The image retrieval method according to claim 13, further comprising: outputting the some or all of the retrieved registered images, by superimposing an image in which the region corresponding to the region of interest of the retrieval image is recognizable. ### Claim 15 (depends on 13) The image retrieval method according to claim 13, further comprising: outputting the some or all of the retrieved registered images, by superimposing an image in which a region corresponding to the region from which the feature of the retrieval image has been extracted is recognizable. ### Claim 16 (depends on 13) The image retrieval method according to claim 13, further comprising: outputting the retrieved registered image and an enlarged image of the region in the retrieved registered image, the region corresponding to the region of interest in the retrieval image. ### Claim 17 (depends on 13) The image retrieval method according to claim 13, further comprising: outputting the some or all of the registered images including the region corresponding to the region of interest of the retrieval image, and the retrieval image. ### Claim 18 (depends on 13) The image retrieval method according to claim 13, wherein the image retrieval method further comprising: storage, in a storage, information associated with each of the plurality of registered images; and outputting the information associated with the retrieved registered image. ### Claim 19 (depends on 13) The image retrieval method according to claim 13, wherein the image retrieval method further comprising: storing registered images captured at a plurality of different timings in a distribution process of the objects to be registered in a storage; retrieving, from the plurality of registered images stored in the storage, a registered image at each of the plurality of timings, the registered image having a feature that matches the extracted feature; and outputting some or all of the retrieved registered images at the plurality of timings in a mode that a region corresponding to the region of interest in the retrieval image is recognizable. ### Claim 20 A non-transitory storage medium storing a program for causing a computer to function as: an acquisition means configured to acquire a retrieval image including a region of interest; an extraction means configured to extract a feature from a region other than the region of interest of the retrieval image; a retrieval means configured to retrieve, from a plurality of registered images obtained by capturing a plurality of objects to be registered and stored in a storage means, a registered image having a feature that matches the feature extracted by the extraction means; and an output means configured to output some or all of the registered images retrieved by the retrieval means in a mode that a region corresponding to the region of interest in the retrieval image is recognizable.