A New Method for Automatic, Real-Time Face Detection and Expression Recognition
University researchers have invented a method for automatic, real-time face detection and expression recognition that is robust for unconstrained situations such as free human motion, varied facial expressions, and many other human and machine factors. The main application has been in human-robot and human-computer interactions, though security applications are also well within reach. Other applications include market surveys, psychological assessment, truth quantification, and automatic tutoring systems.
Patent Number: US7379568B2
Application Number: US2004871494A
Inventor: Movellan, Javier R. | Bartlett, Marian S. | Littlewort, Gwendolen C. | Hershey, John | Fasel, Ian R. | Carlson, Eric C. | Susskind, Josh | Sabe, Kohtaro | Kawamoto, Kenta | Hidai, Kenichi
Priority Date: 24 Jul 2003
Priority Number: US7379568B2
Application Date: 17 Jun 2004
Publication Date: 27 May 2008
IPC Current: G06K000900 | G06T000100 | G06K000962 | G06N000300 | G06T000700
US Class: 382118 | 382159 | 702181 | 706017 | 706025
Assignee Applicant: Sony Corporation,Tokyo | San Diego University of California,Oakland
Title: Weak hypothesis generation apparatus and method, learning apparatus and method, detection apparatus and method, facial expression learning apparatus and method, facial expression recognition apparatus and method, and robot apparatus
Usefulness: Weak hypothesis generation apparatus and method, learning apparatus and method, detection apparatus and method, facial expression learning apparatus and method, facial expression recognition apparatus and method, and robot apparatus
Summary: For generating weak hypothetical data for robot (claimed) such as humanoid robot and four-legged robot used for face expression recognition in security, law-enforcement, psychiatry, education and telecommunication applications.
Novelty: Weak hypothetical data generator for robot for face expression recognition, selects highly-efficient weak postulate or new weak postulate generated by adding preset transformation to highly-efficient weak postulate, for face recognition
機器人/機械
7379568
Tech ID/UC Case 20597/2003-227-0 Related Cases 2003-227-0
美國
