# US 2021/0240993 A1 · Device and method for detecting display of provided credit, and program > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2021/0240993 A1 - Document ID: US20210240993A1 - Status: Published application - Kind: A1 - Application number: 15/734,864 - Filed: 2019-05-13 - Published: 2021-08-05 - Earliest claimed: 2018-06-05 - Assignees: Nippon Telegraph and Telephone Corporation - Applicants: Nippon Telegraph and Telephone Corporation - Inventors: Yasunori OISHI; Takahito KAWANISHI; Kaoru HIRAMATSU; Kunio KASHINO - Web page: https://patentgrep.com/patent/US20210240993A1 ## CPC - H04N 21/44008: {involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream (arrangements characterised by components specially adapted for monitoring, identification or recognition of video in broadcast systems H04H60/59)} - G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting - G06N 3/0464: Convolutional networks [CNN, ConvNet] - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06N 5/04: Inference or reasoning models - G06N 20/00: Machine learning - G06V 10/454: {Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]} - G06V 10/82: using neural networks - G06V 20/46: {Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames} - H04N 17/00: Diagnosis, testing or measuring for television systems or their details - H04N 21/433: Content storage operation, e.g. storage operation in response to a pause request, caching operations - H04N 21/4784: receiving rewards - G06N 3/045: Combinations of networks - G06V 20/44: {Event detection} - G06V 2201/02: Recognising information on displays, dials, clocks - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract The present invention enables detection of a sponsorship credit display in a broadcast program with higher precision. A sponsorship credit display detection device 100 according to the present invention includes: a learning data creation unit 140 that creates, as learning data, from a broadcast program in which the sponsorship credit display has been detected, a still image where the sponsorship credit display is displayed and a still image where the sponsorship credit display is not displayed; a learning unit 150 that uses the learning data created by the learning data creation unit 140 to learn a parameter to be applied to a detection model for detecting a sponsorship credit display in a broadcast program; and a sponsorship credit display detection unit 170 that uses the detection model to which the parameter learned by the learning unit 150 has been applied to detect a sponsorship credit display in a recognition target broadcast program. ## Claims ### Claim 1 A sponsorship credit display detection device for detecting, in a broadcast program, a sponsorship credit display indicating a sponsor of the broadcast program, the device comprising: a learning data creator configured to create, as learning data, from a broadcast program in which the sponsorship credit display has been detected, a still image, wherein the sponsorship credit display is displayed and a still image where the sponsorship credit display is not displayed; a learner configured to, using the learning data created by the learning data creator, to learn a parameter to be applied to a detection model for detecting a sponsorship credit display in a broadcast program; and a sponsorship credit display detector configured to, using the detection model to which the parameter learned by the learner, detect a sponsorship credit display in a recognition target broadcast program. ### Claim 2 (depends on 1) The sponsorship credit display detection device according to claim 1, wherein the sponsorship credit display detector includes: a signal generator configured to: receive still images of the recognition target broadcast program to the detection model in time-series, and, based on an output value of the detection model for each still image, generates a time-series signal indicating in time-series the presence or absence of a sponsorship credit display in each still image; and a time information provider configured to: in a time-series signal generated by the signal generator, detect a segment where still images determined to have a sponsorship credit display continue for at least a predetermined time period as a sponsorship credit display segment where the sponsorship credit display is displayed, and provide time information of the detected sponsorship credit display segment. ### Claim 3 (depends on 1) The sponsorship credit display detection device according to claim 1, wherein the learning data creator is configured to create the learning data for each day of the week from a broadcast program in which the sponsorship credit display has been detected for each day of the week, wherein the learner is configured to use the learning data for each day of the week created by the learning data creator to learn the parameter for each day of the week, and wherein the sponsorship credit display detector is configured to detect the sponsorship credit display in the recognition target broadcast program by applying the parameter learned for the day of the week when the recognition target broadcast program is broadcast to the detection model. ### Claim 4 (depends on 3) The sponsorship credit display detection device according to claim 3, wherein the learner is configured to: learn for each day of the week using latest learning data created by the learning data creator, and update the parameter to be applied to the detection model. ### Claim 5 (depends on 1) The sponsorship credit display detection device according to claim 1, wherein the learning data creator is configured to create, as learning data, from a broadcast program in which the sponsorship credit display has been detected; a still image with the sponsorship credit display and the time when the still image was broadcast, and a still image without the sponsorship credit display and the time when the still image was broadcast. ### Claim 6 A sponsorship credit display detection method in a sponsorship credit display detection device for detecting, in a broadcast program, a sponsorship credit display indicating a sponsor of that broadcast program, the method comprising: creating, as learning data, from a broadcast program in which the sponsorship credit display has been detected, a still image with the sponsorship credit display and a still image without the sponsorship credit display; using the created learning data to learn a parameter to be applied to a detection model for detecting a sponsorship credit display in a broadcast program; and using the detection model to which the learned parameter has been applied to detect a sponsorship credit display in a recognition target broadcast program. ### Claim 7 A computer program causing a computer to function as the sponsorship credit display detection device, the device comprising: a learning data creator configured to create, as learning data, from a broadcast program in which the sponsorship credit display has been detected, a still image, wherein the sponsorship credit display is displayed and a still image where the sponsorship credit display is not displayed; a learner configured to, using the learning data created by the learning data creator, learn a parameter to be applied to a detection model for detecting a sponsorship credit display in a broadcast program; and a sponsorship credit display detector configured to, using the detection model to which the parameter learned by the learner, detect a sponsorship credit display in a recognition target broadcast program. ### Claim 8 (depends on 6) The sponsorship credit display detection method according to claim 6, the method further comprising: receiving still images of the recognition target broadcast program to the detection model in time-series; based on an output value of the detection model for each still image, generating a time-series signal indicating in time-series the presence or absence of a sponsorship credit display in each still image; in a time-series signal generated by the signal generation unit, detecting a segment where still images determined to have a sponsorship credit display continue for at least a predetermined time period as a sponsorship credit display segment where the sponsorship credit display is displayed, and providing time information of the detected sponsorship credit display segment. ### Claim 9 (depends on 6) The sponsorship credit display detection method according to claim 6, the method further comprising: creating the learning data for each day of the week from a broadcast program in which the sponsorship credit display has been detected for each day of the week; learning the parameter for each day of the week by using the learning data for each day of the week; and detecting the sponsorship credit display in the recognition target broadcast program by applying the parameter learned for the day of the week when the recognition target broadcast program is broadcast to the detection model. ### Claim 10 (depends on 6) The sponsorship credit display detection method according to claim 6, the method further comprising: learning for each day of the week using latest learning data; and updating the parameter to be applied to the detection model. ### Claim 11 (depends on 6) The sponsorship credit display detection method according to claim 6, the method further comprising: creating, as learning data, from a broadcast program in which the sponsorship credit display has been detected: a still image with the sponsorship credit display and the time when the still image was broadcast, and a still image without the sponsorship credit display and the time when the still image was broadcast. ### Claim 12 (depends on 7) The computer program according to claim 7 causing the computer to function as the sponsorship credit display detection device, the device comprising: a signal generator configured to: receive still images of the recognition target broadcast program to the detection model in time-series, and, based on an output value of the detection model for each still image, generates a time-series signal indicating in time-series the presence or absence of a sponsorship credit display in each still image; and a time information provider configured to: in a time-series signal generated by the signal generator, detect a segment where still images determined to have a sponsorship credit display continue for at least a predetermined time period as a sponsorship credit display segment where the sponsorship credit display is displayed, and provide time information of the detected sponsorship credit display segment. ### Claim 13 (depends on 7) The computer program according to claim 7 causing the computer to function as the sponsorship credit display detection device, wherein the learning data creator is configured to create the learning data for each day of the week from a broadcast program in which the sponsorship credit display has been detected for each day of the week, wherein the learner is configured to use the learning data for each day of the week created by the learning data creator to learn the parameter for each day of the week, and wherein the sponsorship credit display detector is configured to detect the sponsorship credit display in the recognition target broadcast program by applying the parameter learned for the day of the week when the recognition target broadcast program is broadcast to the detection model. ### Claim 14 (depends on 7) The computer program according to claim 7 causing the computer to function as the sponsorship credit display detection device, wherein the learner is configured to: learn for each day of the week using latest learning data created by the learning data creator, and update the parameter to be applied to the detection model. ### Claim 15 (depends on 7) The computer program according to claim 7 causing the computer to function as the sponsorship credit display detection device, wherein the learning data creator is configured to create, as learning data, from a broadcast program in which the sponsorship credit display has been detected: a still image with the sponsorship credit display and the time when the still image was broadcast, and a still image without the sponsorship credit display and the time when the still image was broadcast. ### Claim 16 (depends on 2) The sponsorship credit display detection device according to claim 2, wherein the learning data creator is configured to create the learning data for each day of the week from a broadcast program in which the sponsorship credit display has been detected for each day of the week, wherein the learner is configured to use the learning data for each day of the week created by the learning data creator to learn the parameter for each day of the week, and wherein the sponsorship credit display detector is configured to detect the sponsorship credit display in the recognition target broadcast program by applying the parameter learned for the day of the week when the recognition target broadcast program is broadcast to the detection model. ### Claim 17 (depends on 2) The sponsorship credit display detection device according to claim 2, wherein the learning data creator is configured to create, as learning data, from a broadcast program in which the sponsorship credit display has been detected: a still image with the sponsorship credit display and the time when the still image was broadcast, and a still image without the sponsorship credit display and the time when the still image was broadcast. ### Claim 18 (depends on 9) The sponsorship credit display detection method according to claim 9, the method further comprising: creating, as learning data, from a broadcast program in which the sponsorship credit display has been detected: a still image with the sponsorship credit display and the time when the still image was broadcast, and a still image without the sponsorship credit display and the time when the still image was broadcast. ### Claim 19 (depends on 10) The sponsorship credit display detection method according to claim 10, the method further comprising: creating, as learning data, from a broadcast program in which the sponsorship credit display has been detected: a still image with the sponsorship credit display and the time when the still image was broadcast, and a still image without the sponsorship credit display and the time when the still image was broadcast. ### Claim 20 (depends on 12) The computer program according to claim 12 causing the computer to function as the sponsorship credit display detection device, wherein the learning data creator is configured to create, as learning data, from a broadcast program in which the sponsorship credit display has been detected: a still image with the sponsorship credit display and the time when the still image was broadcast, and a still image without the sponsorship credit display and the time when the still image was broadcast.