HIDDEN MARKOV MODEL FOR CAMERA HANDOFF
First Claim
Patent Images
1. An integrated method for modeling the handoff between cameras for tracking a specific individual, comprising:
- creating a representation of overlaps, gaps, and allowable movement among fields of view of the cameras, wherein the representation is modeled as states in a Hidden Markov Model (HMM);
training the HMM using video of people walking through the fields of view of the cameras;
selecting a person to be tracked; and
identifying a best camera area using the HMM.
1 Assignment
0 Petitions
Accused Products
Abstract
An integrated method for modeling the handoff between cameras for tracking a specific individual, including: creating a representation of overlaps, gaps, and allowable movement among the fields of view of the cameras, wherein the representation is modeled as states in a Hidden Markov Model (HMM); training the HMM using video of people walking through the fields of view of the cameras; selecting a person to be tracked; and identifying the best camera area using the HMM.
-
Citations
22 Claims
-
1. An integrated method for modeling the handoff between cameras for tracking a specific individual, comprising:
-
creating a representation of overlaps, gaps, and allowable movement among fields of view of the cameras, wherein the representation is modeled as states in a Hidden Markov Model (HMM); training the HMM using video of people walking through the fields of view of the cameras; selecting a person to be tracked; and identifying a best camera area using the HMM. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. A program of instructions executable by a computer to perform a function for modeling the handoff between cameras for tracking a specific individual, the function comprising the steps of:
-
creating a representation of overlaps, gaps, and allowable movement among the fields of view of the cameras, wherein the representation is modeled as states in a Hidden Markov Model (HMM); training the HMM using video of people walking through the fields of view of the cameras; selecting a Very Important Person (VIP) to be tracked; identifying the best camera area using the HMM; and identifying the best camera in the camera area using the HMM. - View Dependent Claims (17, 18, 19, 20, 21, 22)
-
Specification