公开(公告)号：US20170140546A1

公开(公告)日：2017-05-18

申请号：US15420570

申请日：2017-01-31

Applicant: Mitsubishi Electric Research Laboratories, Inc.

Inventor： Hassan Mansour ,  Caglayan Dicle ,  Dong Tian ,  Mouhacine Benosman ,  Anthony Vetro

IPC: G06T7/215 ,  G06T7/136 ,  G06T5/00 ,  G06T7/174 ,  G06T7/246 ,  G06K9/00

CPC classification number: G06T7/215 ,  G06K9/00335 ,  G06K9/00718 ,  G06K9/00765 ,  G06K9/00778 ,  G06K9/00785 ,  G06K9/52 ,  G06T1/60 ,  G06T5/002 ,  G06T7/136 ,  G06T7/174 ,  G06T7/20 ,  G06T7/246 ,  G06T2207/10016 ,  G06T2207/30196 ,  G06T2207/30232 ,  H04N19/176 ,  H04N19/426

Abstract: A method and system determines flows by first acquiring a video of the flows with a camera, wherein the flows are pedestrians in a scene, wherein the video includes a set of frames. Motion vectors are extracted from each frame in the set, and a data matrix is constructed from the motion vectors in the set of frames. A low rank Koopman operator is determined from the data matrix and a spectrum of the low rank Koopman operator is analyzed to determine a set of Koopman modes. Then, the frames are segmented into independent flows according to a clustering of the Koopman modes.

公开(公告)号：US09892520B2

公开(公告)日：2018-02-13

申请号：US15420570

申请日：2017-01-31

Applicant: Mitsubishi Electric Research Laboratories, Inc.

Inventor： Hassan Mansour ,  Caglayan Dicle ,  Dong Tian ,  Mouhacine Benosman ,  Anthony Vetro

IPC: G06K9/34 ,  G06T7/215 ,  G06T7/246 ,  G06K9/00 ,  G06T5/00 ,  G06T7/174 ,  G06T7/136

CPC classification number: G06T7/215 ,  G06K9/00335 ,  G06K9/00718 ,  G06K9/00765 ,  G06K9/00778 ,  G06K9/00785 ,  G06K9/52 ,  G06T1/60 ,  G06T5/002 ,  G06T7/136 ,  G06T7/174 ,  G06T7/20 ,  G06T7/246 ,  G06T2207/10016 ,  G06T2207/30196 ,  G06T2207/30232 ,  H04N19/176 ,  H04N19/426

Abstract: A method and system determines flows by first acquiring a video of the flows with a camera, wherein the flows are pedestrians in a scene, wherein the video includes a set of frames. Motion vectors are extracted from each frame in the set, and a data matrix is constructed from the motion vectors in the set of frames. A low rank Koopman operator is determined from the data matrix and a spectrum of the low rank Koopman operator is analyzed to determine a set of Koopman modes. Then, the frames are segmented into independent flows according to a clustering of the Koopman modes.

公开(公告)号：US09600896B1

公开(公告)日：2017-03-21

申请号：US14932113

申请日：2015-11-04

Applicant: Mitsubishi Electric Research Laboratories, Inc.

Inventor： Hassan Mansour ,  Caglayan Dicle ,  Dong Tian ,  Mouhacine Benosman ,  Anthony Vetro

IPC: G06K9/34 ,  G06T7/00 ,  G06K9/00 ,  G06T7/20 ,  G06K9/52 ,  G06T1/60 ,  H04N19/176 ,  H04N19/426

CPC classification number: G06T7/215 ,  G06K9/00335 ,  G06K9/00718 ,  G06K9/00765 ,  G06K9/00778 ,  G06K9/00785 ,  G06K9/52 ,  G06T1/60 ,  G06T5/002 ,  G06T7/136 ,  G06T7/174 ,  G06T7/20 ,  G06T7/246 ,  G06T2207/10016 ,  G06T2207/30196 ,  G06T2207/30232 ,  H04N19/176 ,  H04N19/426

Abstract: A method and system determines flows by first acquiring a video of the flows with a camera, wherein the flows are pedestrians in a scene, wherein the video includes a set of frames. Motion vectors are extracted from each frame in the set, and a data matrix is constructed from the motion vectors in the set of frames. A low rank Koopman operator is determined from the data matrix and a spectrum of the low rank Koopman operator is analyzed to determine a set of Koopman modes. Then, the frames are segmented into independent flows according to a clustering of the Koopman modes.