-
Patent Agency Ranking
公开(公告)号：JP2003168106A
公开(公告)日：2003-06-13
申请号：JP2001369076
申请日：2001-12-03
Applicant: ATR ADVANCED TELECOMM RES INST
Inventor： KAWATO SHINJIRO , TETSUYA SHINJI
Abstract: PROBLEM TO BE SOLVED: To provide a method for tracking eye positions by which the eye positions can be tracked from a general facial image with high robustness and without unduly constraining a user. SOLUTION: The method for tracking the eye positions includes a step 242 for acquiring an image, a step 244 for predicting a middle forehead position of the facial image in the image, a step 246 for searching a point conforming to an image pattern of a middle forehead best in the vicinity of the predicted middle forehead position, a step 248 for predicting positions of both eyes in the image based on a position of the searched point and relative positions of both eyes and for searching a point to be the center of each of two areas satisfying conditions predetermined centering around the predicted areas and a step 254 for updating the middle forehead position, the relative positions of both eyes and the image pattern of the middle forehead. COPYRIGHT: (C)2003,JPO
-
公开(公告)号：JP2004012923A
公开(公告)日：2004-01-15
申请号：JP2002167577
申请日：2002-06-07
Applicant: ATR ADVANCED TELECOMM RES INST
Inventor： NOMA HARUO , TETSUYA SHINJI
Abstract: PROBLEM TO BE SOLVED: To provide a learning system by which a user can experience a program. SOLUTION: The learning system 100 is provided with a freighter 10, a baggage operation unit 20, baggages 30, an engine 40, a signal unit 60, a point unit 70, a baggage site 80, an input unit 90, and an output unit 110. The freighter 10, the baggage operation unit 20, the baggages 30, the engine 40, the signal unit 60, the point unit 70, the baggage site 80, the input unit 90, and the output unit 110 are suitably selected to constitute a railroad model expressing a program for performing various operation. COPYRIGHT: (C)2004,JPO
-
公开(公告)号：JP2003233798A
公开(公告)日：2003-08-22
申请号：JP2002030809
申请日：2002-02-07
Applicant: ATR ADVANCED TELECOMM RES INST
Inventor： HAGITA NORIHIRO , MASE KENJI , TADENUMA MAKOTO , TETSUYA SHINJI , KATAGIRI TAKAHIRO
Abstract: PROBLEM TO BE SOLVED: To provide an experience transmission system making it possible to mutually share experiences of human beings. SOLUTION: A cooperation media 1a acquires and stores experience information of a first user by automatically and cooperatively interacting with the first user. A cooperation media 2b makes a second user to follow the experiment of the first user using the experiment information of the first user read from the cooperation media 1a via a network 3. COPYRIGHT: (C)2003,JPO
-
公开(公告)号：JP2003168121A
公开(公告)日：2003-06-13
申请号：JP2001369075
申请日：2001-12-03
Applicant: ATR ADVANCED TELECOMM RES INST
Inventor： KAWATO SHINJIRO , TETSUYA SHINJI
Abstract: PROBLEM TO BE SOLVED: To provide a method for extracting eyes by which eye positions can be detected from a general face image without unduly constraining a user. SOLUTION: This method includes: a step 122 for calculating a differential image between two images to be outputted by a camera; a step 132 for cancelling a differential area due to transfer of a subject between the two images in the differential image and for extracting the remaining differential area based on the differential image and the two images; a step for selecting two candidate points for eyes from the extracted remaining differential area according to predetermined conditions; and a step for determining whether or not the two selected candidate points for the eyes satisfy the predetermined conditions and for extracting the two candidate points for the eyes as the eye positions when the predetermined conditions are satisfied. COPYRIGHT: (C)2003,JPO
-
公开(公告)号：JP2003178303A
公开(公告)日：2003-06-27
申请号：JP2001376000
申请日：2001-12-10
Applicant: ATR ADVANCED TELECOMM RES INST
Inventor： UCHIUMI AKIRA , TETSUYA SHINJI
Abstract: PROBLEM TO BE SOLVED: To provide an object recognition device recognizing an object on the basis of a characteristic of appearance of the object itself and not by extracting the target object as a moving area or a particular color area. SOLUTION: A discrimination axis is determined by extracting geometrical structures of a background image and a person image picked up by a camera as a difference of picture element distributions and constructing a statistical model of appearance of the person image. The model and an inputted image are compared and a score is calculated on the basis of the discrimination axis of the model and a distance value vector of the inputted image. A recognition process is efficiently carried out by a small number of models by judging that a person has been detected when the score exceeds a threshold. COPYRIGHT: (C)2003,JPO
-
-
-
-
Search Results
5
records
(took 0.03 seconds)
