发明名称 System and method for identifying a user through an object held in a hand
摘要 A system and method for identifying a user through an object held by a hand of the user, according to an image of the skin surface print of a portion of the hand of the user, which is optionally the skin surface print of at least a portion of the hand between the metacarpophalangeal joint and a distal interphalangeal joint of one or more fingers. Optionally, the image only includes the skin surface print of at least a portion of the hand between the metacarpophalangeal joint and a distal interphalangeal joint of one or more fingers. The method for identifying the user may also optionally only use selected portions of this image as described herein. Related apparatus and methods are also described.
申请公布号 US8831298(B2) 申请公布日期 2014.09.09
申请号 US201113805576 申请日期 2011.06.20
申请人 Cisco Technology Inc. 发明人 Farkash Eyal
分类号 G06K9/00;G06K9/46 主分类号 G06K9/00
代理机构 Husch Blackwell LLP 代理人 Husch Blackwell LLP
主权项 1. A method for identifying a person holding an object with a hand, wherein the object comprises an image capture device and a transmitter, the method comprising: providing a remote processing device comprising a receiver and a database, wherein the transmitter of the object and said receiver of said remote processing device are in communication and wherein said remote processing device and the object are physically separate; capturing an image of at least a portion of the hand, said portion of the hand comprising at least a portion of the hand between a metacarpophalangeal joint and a distal interphalangeal joint of one or more fingers, said image featuring only a skin surface print; sending said image with said transmitter to said receiver; constructing a signature from data from said image by said remote processing device, said constructing said signature comprising converting said image to luminance levels to form a luminance matrix and converting said luminance matrix to form a vector through a lossy conversion process, wherein said signature comprises said vector, and said converting said luminance matrix to said vector comprises: dividing said luminance matrix into a plurality of blocks each of N by N pixels, wherein N refers to a number of pixels in each row and in each column of each block of said luminance matrix; andapplying said lossy conversion process to at least some of said plurality of blocks, said lossy conversion process comprising a transform selected from the group consisting of a DCT (discrete cosine transform) transform, and a DFT (discrete Fourier transform), and also comprising applying a filter after applying said transform, said applying said filter comprising multiplying by a power vertical vector; and searching through said database for a matching signature to said Signature obtained from said image to identify the person.
地址 San Jose CA US