发明名称 AUTOMATIC FACE DISCOVERY AND RECOGNITION FOR VIDEO CONTENT ANALYSIS
摘要 Systems and methods are provided for automatically classifying videos based on faces discovered in the videos, wherein the discovered faces are not known to be associated with a particular category of videos. The detected face is compared to a set of unknown faces to generate a cluster of unknown faces that each match with the detected face. A set of categorized videos is identified based on the cluster of unknown faces. One or more categories are assigned to the video based on categories from the set of categorized videos so that the video can be automatically classified based on the detected face even though the detected face is not associated with a known person.
申请公布号 US2014375886(A1) 申请公布日期 2014.12.25
申请号 US201414308181 申请日期 2014.06.18
申请人 Set Media, Inc. 发明人 GALLEGUILLOS Carolina;SHAH Hardik Jayant;BALAGOPALAN Arun
分类号 G06F17/30 主分类号 G06F17/30
代理机构 代理人
主权项 1. A computerized method for automatically classifying videos based on faces discovered in the videos, wherein the discovered faces are not known to be associated with a particular category of videos, the method comprising: receiving, by a computing device, a video for classification; detecting, by the computing device, a face of a person in the video; determining, by the computing device, that the detected face is not associated with a set of known people stored in a known person database that is in communication with the computing device, wherein each known person from the set of known people is associated with one or more categories that can be used to classify videos in which the known person appears; comparing, by the computing device, the detected face to a set of unknown faces in an unknown face database in communication with the computing device, to generate a cluster of unknown faces that each match with the detected face by meeting a predetermined match threshold; identifying, by the computing device, a set of categorized videos based on the cluster of unknown faces; and assigning, by the computing device, one or more categories to the video based on categories from the set of categorized videos so that the video can be automatically classified based on the detected face even though the detected face is not associated with a known person.
地址 San Francisco CA US
您可能感兴趣的专利