3D human pose and shape modeling
First Claim
Patent Images
1. A method of detecting human objects in a video, comprising:
- determining pixels of a video image are foreground pixels, the group of foreground pixels constituting a foreground blob set of one or more foreground blobs;
for each of N locations within the video image, where N is an integer, comparing a corresponding predetermined shape with the foreground blob set to obtain a corresponding probability of a human at the location, thereby obtaining N probabilities corresponding to the N locations;
using the N probabilities, determining X humans are represented by the foreground blob set, where X is whole number;
using the determination of the representation of X humans, determining a crowd density within a first area of the real world;
comparing the crowd density to a threshold; and
providing at least one of a report, an alarm, and an event detection when the crowd density exceeds the threshold,wherein a size of the corresponding predetermined shape for each of the N locations is determined in response to calibration of a video system,wherein the video system is used to obtain the video image.
5 Assignments
0 Petitions
Accused Products
Abstract
Methods, devices and systems for performing video content analysis to detect humans or other objects of interest a video image is disclosed. The detection of humans may be used to count a number of humans, to determine a location of each human and/or perform crowd analyses of monitored areas.
-
Citations
28 Claims
-
1. A method of detecting human objects in a video, comprising:
-
determining pixels of a video image are foreground pixels, the group of foreground pixels constituting a foreground blob set of one or more foreground blobs; for each of N locations within the video image, where N is an integer, comparing a corresponding predetermined shape with the foreground blob set to obtain a corresponding probability of a human at the location, thereby obtaining N probabilities corresponding to the N locations; using the N probabilities, determining X humans are represented by the foreground blob set, where X is whole number; using the determination of the representation of X humans, determining a crowd density within a first area of the real world; comparing the crowd density to a threshold; and providing at least one of a report, an alarm, and an event detection when the crowd density exceeds the threshold, wherein a size of the corresponding predetermined shape for each of the N locations is determined in response to calibration of a video system, wherein the video system is used to obtain the video image. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21)
-
-
22. A method of detecting human objects in a video, comprising:
-
determining pixels of a video image of a real world scene are foreground pixels, the group of foreground pixels constituting a foreground blob set of one or more foreground blobs; for each of N locations within the video image, where N is an integer, comparing a corresponding predetermined shape with the foreground blob set to determine X humans are represented by the foreground blob set, where X is whole number and a location of each of the X humans is determined as a location within a horizontal plane of the real world; using the determination of the representation of X humans, determining a crowd density within a first area of the real world; comparing the crowd density to a threshold; and providing at least one of a report, an alarm, and an event detection when the crowd density exceeds the threshold, wherein a size of the corresponding predetermined shape for each of the N locations is determined in response to calibration of a video system, wherein the video system is used to obtain the video image. - View Dependent Claims (23, 24, 25, 26, 27, 28)
-
Specification