Multi-cue object association

Patent metadata

Field	Value
Publication number	US-9104919-B2
Application number	US-201414506974-A
Country	US
Kind code	B2
Filing date	Oct 6, 2014
Priority date	Oct 5, 2012
Publication date	Aug 11, 2015
Grant date	Aug 11, 2015

Abstract

Multiple discrete objects within a scene image captured by a single camera track are distinguished as un-labeled from a background model within a first frame of a video data input. Object position, object appearance and/or object size attributes are determined for each of the blobs, and costs determined to assign to existing blobs of existing object tracks as a function of the determined attributes. The un-labeled object blob that has a lowest cost of association with any of the existing object tracks is labeled with the label of that track having the lowest cost, said track is removed from consideration for labeling remaining un-labeled object blobs, and the process iteratively repeated until each of the track labels have been used to label one of the un-labeled blobs.

First claim

What is claimed is:

1. A method to track multiple discrete objects within a scene image captured by a single camera, the method comprising:
distinguishing a plurality of foreground object bounding box blobs from a background model within a first frame of a video data input from a single camera view via a background subtraction process;
initializing each of the plurality of foreground object bounding box blobs as un-labeled;
determining at least one image attribute for each of…

Assignees

• IBM

Inventors

• Datta Ankur
• Feris Rogerio S
• Pankanti Sharathchandra U
• Zhai Yun

Classifications

• G06V20/53

  Physics · mapped topic

• G06T7/248 Primary

  Physics · mapped topic

• G06V10/28

  Physics · mapped topic

• G06T2207/10016

  Physics · mapped topic

• G06T7/2053

  Physics · mapped topic