公开(公告)号：US20150347903A1

公开(公告)日：2015-12-03

申请号：US14291910

申请日：2014-05-30

Applicant: ADOBE SYSTEMS INCORPORATED

Inventor： Neha Saxena ,  Pulkit Agarwal ,  Raman Gupta

IPC: G06N5/04 ,  G06T11/20

CPC classification number: G06T11/206

Abstract: A computer implemented method and apparatus for acquiring sentiment analysis information based on user reactions to displayable content. The method comprises receiving reaction data of at least one individual user viewing portions of displayable content, wherein the reaction data of each user includes indications representative of at least one of a time-varying emotional state of the user and a time-varying level of interest of the user captured during viewing of the displayable content; and performing sentiment analysis on the reaction data by at least one of: categorizing portions of the displayable content based on one of the reaction of one user or aggregated reactions of a plurality of users, and identifying at least one portion of the displayable content having one of a more favorable and a less favorable reaction by the at least one user, by comparison to one of a baseline and reactions to other portions of the displayable content.

Abstract translation: 一种基于用户对可显示内容的反应来获取情绪分析信息的计算机实现的方法和装置。 该方法包括接收至少一个单独用户观看可显示内容的部分的反应数据，其中每个用户的反应数据包括代表用户的时变感情状态和时变感兴趣水平中的至少一个的指示 在观看可显示内容期间捕获的用户; 以及基于一个用户的反应或多个用户的聚合反应中的一个对所述可显示内容的部分进行分类并且将所述可显示内容的至少一部分识别为具有一个用户的可显示内容的至少一部分来对所述反应数据进行情绪分析 通过与可显示内容的其他部分的基线和反应之一进行比较，由至少一个用户进行更有利和更不利的反应。

公开(公告)号：US20170195746A1

公开(公告)日：2017-07-06

申请号：US14988059

申请日：2016-01-05

Applicant: Adobe Systems Incorporated

Inventor： Raman Gupta

IPC: H04N21/81 ,  H04N21/262 ,  H04N21/6587

CPC classification number: H04N21/812 ,  H04N21/44222 ,  H04N21/47217 ,  H04N21/6547 ,  H04N21/6582

Abstract: A digital medium environment is described to control a start time at which a skippable video advertisement made available by an advertisement provider begins playback at a client. A skippable video advertisement is selected for playback by the client. Data is queried that identifies a skip time corresponding to a time within the skippable video advertisement at which the skippable video advertisement was skipped during a previous viewing of the skippable video advertisement by a user of the client. One of a plurality of start times associated with the skippable video advertisement is then selected based at least in part on the skip time. The skippable video advertisement and the selected start time are then provided to the client effective to cause the skippable video advertisement to begin playback at the selected start time.

公开(公告)号：US09532106B1

公开(公告)日：2016-12-27

申请号：US14809649

申请日：2015-07-27

Applicant: Adobe Systems Incorporated

Inventor： Raman Gupta ,  Neha Saxena ,  Anuj Jain

IPC: H04N5/445 ,  H04N21/475 ,  H04N21/422 ,  H04N21/258 ,  H04N21/25 ,  H04N21/442 ,  H04N21/466 ,  H04N21/2668 ,  H04N21/81

CPC classification number: H04N21/4755 ,  H04N21/251 ,  H04N21/25883 ,  H04N21/25891 ,  H04N21/2668 ,  H04N21/42203 ,  H04N21/44222 ,  H04N21/4667 ,  H04N21/4668 ,  H04N21/4756 ,  H04N21/812 ,  H04N21/8133 ,  H04N21/8456

Abstract: Video character-based content targeting is described. In one or more embodiments, users make selections of characters in video content in conjunction with playback of the video content. For example, a user is prompted during playback of video content to select a character in the video content that the user likes. An indication is received of selections made by the user of one or more characters, each of which is associated with information that describes a degree to which the character exhibits a plurality of personality attributes. Once received, the user's selections of the one or more characters from the video content are analyzed. In particular, the selections are analyzed to ascertain a degree to which the user identifies with the plurality of personality attributes. The degree to which the user is determined to identify with the plurality of personality attributes is then used to control communication of content to the user.

Abstract translation: 描述基于视频字符的内容定向。 在一个或多个实施例中，用户结合视频内容的回放来选择视频内容中的字符。 例如，在播放视频内容期间提示用户选择用户喜欢的视频内容中的角色。 接收到用户对一个或多个字符进行的选择的指示，每个字符与描述角色表现多个个性属性的程度的信息相关联。 一旦接收到，就分析用户对视频内容中的一个或多个字符的选择。 特别地，分析选择以确定用户用多个个性属性识别的程度。 然后使用用户被确定与多个个性属性一起识别的程度来控制内容与用户的通信。

公开(公告)号：US20170372360A1

公开(公告)日：2017-12-28

申请号：US15193183

申请日：2016-06-27

Applicant: Adobe Systems Incorporated

Inventor： Ashish Duggal ,  Niraj Gupta ,  Piyush Gupta ,  Raman Gupta

IPC: G06Q30/02 ,  H04N13/02 ,  G06T19/00 ,  H04N13/04 ,  G06F3/01

CPC classification number: G06Q30/0255 ,  G11B27/00 ,  H04N13/282 ,  H04N13/368

Abstract: Audience segmentation can be based on a viewing angle of a user viewing a video of a multi-angle viewing environment. During playback, a sequence of the user-controlled viewing angles of the video are recorded. The sequence represents the viewing angle of the user at a given point in time. Based on the sequences of several users, a predominant sequence of viewing angles of the video is determined. One or more audience segment tags are assigned to the predominant sequence of viewing angles. During subsequent playbacks of the video, the sequence(s) of user-controlled viewing angles of the video are recorded. The recorded sequence(s) of the subsequent user(s) are compared to the predominant sequence of viewing angles of the video, and the subsequent user(s) are assigned to an audience segment based on the comparison and the corresponding audience segment tags.

公开(公告)号：US09536329B2

公开(公告)日：2017-01-03

申请号：US14291910

申请日：2014-05-30

Applicant: Adobe Systems Incorporated

Inventor： Neha Saxena ,  Pulkit Agarwal ,  Raman Gupta

IPC: G06F17/00 ,  G06F17/20 ,  G06T11/20

CPC classification number: G06T11/206

Abstract: A computer implemented method and apparatus for acquiring sentiment analysis information based on user reactions to displayable content. The method comprises receiving reaction data of at least one individual user viewing portions of displayable content, wherein the reaction data of each user includes indications representative of at least one of a time-varying emotional state of the user and a time-varying level of interest of the user captured during viewing of the displayable content; and performing sentiment analysis on the reaction data by at least one of: categorizing portions of the displayable content based on one of the reaction of one user or aggregated reactions of a plurality of users, and identifying at least one portion of the displayable content having one of a more favorable and a less favorable reaction by the at least one user, by comparison to one of a baseline and reactions to other portions of the displayable content.

Abstract translation: 一种基于用户对可显示内容的反应来获取情绪分析信息的计算机实现的方法和装置。 该方法包括接收至少一个单独用户观看可显示内容的部分的反应数据，其中每个用户的反应数据包括代表用户的时变感情状态和时变感兴趣水平中的至少一个的指示 在观看可显示内容期间捕获的用户; 以及基于一个用户的反应或多个用户的聚合反应中的一个对所述可显示内容的部分进行分类并且将所述可显示内容的至少一部分识别为具有一个用户的可显示内容的至少一部分来对所述反应数据进行情绪分析 通过与可显示内容的其他部分的基线和反应之一进行比较，由至少一个用户进行更有利和更不利的反应。