发明名称 Face recognition system and method for taking group photo
摘要 A face recognition method recognizes complete faces in a preview image of a group photo taken by an image capturing device, divides the preview image into default blocks if the number of the complete faces is no more than a predetermined number, and adjusts the default blocks into adjusted blocks, wherein the adjusted blocks comprising complete face blocks and non-complete face blocks, wherein each of the complete face blocks contains only one complete face while each of the non-complete face blocks does not contain any complete face, and one of the most adjacent complete face contains one of the most adjacent complete faces. The method analyzes the adjusted blocks to obtain target blocks from the non-complete face blocks which is most adjacent to any one of the complete face blocks and recognizes facial features in each of the target blocks and obtain non-complete face blocks.
申请公布号 US8902323(B2) 申请公布日期 2014.12.02
申请号 US201213568149 申请日期 2012.08.07
申请人 Fu Tai Hua Industry (Shenzhen) Co., Ltd.;Hon Hai Precision Industry Co., Ltd. 发明人 Zhang Pan
分类号 H04N5/228 主分类号 H04N5/228
代理机构 Novak Druce Connolly Bove + Quigg LLP 代理人 Novak Druce Connolly Bove + Quigg LLP
主权项 1. A face recognition method implemented on an image capturing device for taking group photos, comprising: obtaining one preview image; recognizing complete faces in the preview image; dividing the preview image into a plurality of default blocks if the number of the complete faces is more than or equal to a predetermined number; adjusting the default blocks into a plurality of adjusted blocks, the adjusted blocks comprising a number of complete face blocks and a number of non-complete face blocks, wherein each of the complete face blocks contains only one complete face while each of the non-complete face blocks does not contains any complete face, and one of the most adjacent complete face blocks contains one of the most adjacent complete face; analyzing the adjusted blocks to obtain a plurality of target blocks from the non-complete face blocks which is most adjacent to any one of the complete face blocks; and recognizing facial features in each of the target blocks and obtain a plurality of non-complete face blocks, each of which comprises any of the facial features.
地址 Shenzhen CN