# US 9,521,443 B2 · Method and apparatus for generating quality estimators > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 9,521,443 B2 - Document ID: US9521443B2 - Status: Granted - Kind: B2 - Application number: 14/658,746 - Filed: 2015-03-16 - Published: 2015-07-02 - Granted: 2016-12-13 - Earliest claimed: 2013-09-13 - Assignees: AT&T Intellectual Property I, L.P. - Applicants: AT&T Intellectual Property I, LP - Inventors: Amy Ruth Reibman; Kenneth Shirley; Chao Tian - Web page: https://patentgrep.com/patent/US9521443B2 ## CPC - G06V 10/993: {Evaluation of the quality of the acquired pattern} - G06F 18/22: Matching criteria, e.g. proximity measures - G06T 7/0002: {Inspection of images, e.g. flaw detection} - G06V 10/30: Noise filtering - H04N 21/23418: {involving operations for analysing video streams, e.g. detecting features or characteristics (television picture signal circuitry for scene change detection H04N5/147; filtering for image enhancement G06T5/00; methods or arrangements for recognising scenes G06V20/00; arrangements characterised by components specially adapted for monitoring, identification or recognition of video in broadcast systems H04H60/59)} - H04N 21/23439: {for generating different versions} - H04N 21/252: {Processing of multiple end-users' preferences to derive collaborative data} - H04N 21/25891: {being end-user preferences (retrieval of video data in a video database based on user preferences G06F16/739; arrangements for recognizing users' preferences H04H60/46; user profiles in network data switching protocols H04L67/306; processing of user preferences or user profiles in wireless networks H04W8/18)} - H04N 21/266: Channel or content management, e.g. generation and management of keys and entitlement messages in a conditional access system, merging a VOD unicast channel into a multicast channel - H04N 21/64322: {IP} - G06T 2207/30168: Image quality inspection - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION ## Abstract A system that incorporates teachings of the present disclosure may include, for example, sampling a variable effect distribution of viewing preference data to determine a first set of effects comprising a plurality of first distortion type effects associated with a first distortion type of a first image and to determine a second set of effects comprising a plurality of second distortion type effects associated with the second distortion type of a second image, calculating a preference estimate from a logistic regression model of the viewing preference data according to the first set of effects and the second set of effects, wherein the preference estimate comprises a probability that the first image is preferred over the second image, and selecting one of the first distortion type or the second distortion type according to the preference estimate. Other embodiments are disclosed. ## Claims ### Claim 1 A method comprising: obtaining, by a processing system comprising a processor, a first set of viewer bias data for a first viewer of first image content that has been distorted by a first distortion type and a second set of viewer bias data for a second viewer of second image content that has been distorted by a second distortion type; generating, by the processing system, a regression model of viewing preference data that is based on the first set of viewer bias data and the second set of viewer bias data; calculating, by the processing system, a preference estimate from the regression model of viewing preference data, wherein the preference estimate corresponds to a probability that the first image content that has been distorted by the first distortion type is preferred by a third over the second image content that has been distorted by the second distortion type; selecting, by the processing system, a target distortion type from the first distortion type or the second distortion type according to the preference estimate; and adjusting, by the processing system, content for distribution to viewer equipment utilizing a distribution process that is selected based on the target distortion type. ### Claim 2 (depends on 1) The method of claim 1, wherein the regression model of viewing preference data comprises a logistic regression model of the viewing preference data. ### Claim 3 (depends on 1) The method of claim 1, wherein the distribution process comprises a target video compression process selected from among a group of video compression processes. ### Claim 4 (depends on 1) The method of claim 1, wherein the first distortion type comprises one of a Gaussian blur, an image compression, an additive Gaussian noise, or a combination thereof. ### Claim 5 (depends on 1) The method of claim 1, wherein the first set of viewer bias data and the second set of viewer bias data are obtained by sampling a variable effect distribution of the viewing preference data associated with the first viewer and the second viewer. ### Claim 6 (depends on 1) The method of claim 1, wherein the first image content originates from first media content and the second image content originates from second media content, and wherein the first media content is different from the second media content. ### Claim 7 (depends on 1) The method of claim 1, further comprising calculating, by the processing system, a confidence value according to the first set of viewer bias data and the second set of viewer bias data, wherein the confidence value further defines a characteristic of the preference estimate. ### Claim 8 (depends on 7) The method of claim 7, wherein the selecting of the target distortion type is further according to the confidence value. ### Claim 9 A computer-readable storage device comprising executable instructions, which when executed by a processing system including a processor cause the processing system to perform operations comprising: logging image pair preference observations from a first viewer of first image content that has been distorted by a first distortion type and a second set of viewer bias data for a second viewer of second image content that has been distorted by a second distortion type; sampling a variable effect distribution constructed from the image pair preference observations to obtain a first set of viewer bias data for the first viewer and a second set of viewer bias data for the second viewer; generating a regression model of viewing preference data that is based on the first set of viewer bias data and the second set of viewer bias data; calculating a preference estimate from the regression model of viewing preference data, wherein the preference estimate corresponds to a probability that the first image content that has been distorted by the first distortion type is preferred by a third viewer over the second image content that has been distorted by the second distortion type; selecting a target distortion type of the first distortion type or the second distortion type for processing content according to the preference estimate; and adjusting content for distribution to viewer equipment utilizing a distribution process that is selected based on the target distortion type. ### Claim 10 (depends on 9) The storage device of claim 9, wherein the regression model of viewing preference data comprises a logistic regression model. ### Claim 11 (depends on 9) The storage device of claim 9, wherein the operations further comprise determining a confidence value according to the first set of viewer bias data and the second set of viewer bias data, wherein the confidence value further defines a characteristic of the preference estimate. ### Claim 12 A server comprising: a memory having executable instructions; and a processing system including a processor coupled with the memory, wherein the processing system when executing the executable instructions performs operations comprising: receiving content for distribution; calculating a preference estimate from a regression model of viewing preference data, wherein the regression model is based on a first set of viewer bias data for a first viewer of first image content that has been altered by a first distortion type and a second set of viewer bias data for a second viewer of second image content that has been altered by a second distortion type, wherein the preference estimate corresponds to a probability that the first image content is preferred over the second image content; selecting a target distortion type from the first distortion type or the second distortion type according to the preference estimate; and adjusting the content for distribution to viewer equipment utilizing a distribution process that is selected based on the target distortion type. ### Claim 13 (depends on 12) The server of claim 12, wherein the regression model of the viewing preference data comprises a logistic regression model. ### Claim 14 (depends on 12) The server of claim 12, wherein the first set of viewer bias data and the second set of viewer bias data are obtained by sampling a variable effect distribution of the viewing preference data associated with the first viewer and the second viewer. ### Claim 15 (depends on 12) The server of claim 12, wherein the distribution process comprises a target video compression process selected from among a group of video compression processes. ### Claim 16 (depends on 12) The server of claim 12, wherein the first image content originates from first media content and the second image content originates from second media content, and wherein the first media content is different from the second media content. ### Claim 17 (depends on 12) The server of claim 12, wherein the operations further comprise determining a confidence value according to the first set of viewer bias data and the second set of viewer bias data, wherein the confidence value further defines a characteristic of the preference estimate. ### Claim 18 (depends on 17) The server of claim 17, wherein the selecting of the target distortion type is further according to the confidence value. ### Claim 19 (depends on 12) The server of claim 12, wherein the content comprises video content. ### Claim 20 (depends on 12) The server of claim 12, wherein the first distortion type comprises one of a Gaussian blur, an image compression, an additive Gaussian noise, or a combination thereof.