Methods, devices and systems for detecting objects in a video
First Claim
Patent Images
1. A method of detecting human objects in a video, comprising:
- determining pixels of a video image are foreground pixels, a group of the foreground pixels constituting a foreground blob set of one or more foreground blobs;
for each of N predetermined shapes at corresponding ones of N predetermined locations within the video image, where N is a positive integer greater than one, calculating a first value by comparing an overlap of the corresponding predetermined shape with the foreground blob set, the first value being used to obtain a corresponding probability of a human at the corresponding predetermined 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 a whole number; and
providing at least one of a report, an alarm, and an event detection using the determination of the representation of X humans,wherein a size of the corresponding predetermined shape for each of the N locations is determined in response to calibration of the video system, andwherein the comparing of the corresponding predetermined shape with the foreground blob set for each of the N predetermined shapes comprises analyzing an amount of an overlapping area of the corresponding predetermined shape with the foreground blob set.
3 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 analyzes of monitored areas.
-
Citations
29 Claims
-
1. A method of detecting human objects in a video, comprising:
-
determining pixels of a video image are foreground pixels, a group of the foreground pixels constituting a foreground blob set of one or more foreground blobs; for each of N predetermined shapes at corresponding ones of N predetermined locations within the video image, where N is a positive integer greater than one, calculating a first value by comparing an overlap of the corresponding predetermined shape with the foreground blob set, the first value being used to obtain a corresponding probability of a human at the corresponding predetermined 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 a whole number; and providing at least one of a report, an alarm, and an event detection using the determination of the representation of X humans, wherein a size of the corresponding predetermined shape for each of the N locations is determined in response to calibration of the video system, and wherein the comparing of the corresponding predetermined shape with the foreground blob set for each of the N predetermined shapes comprises analyzing an amount of an overlapping area of the corresponding predetermined shape with the foreground blob set. - 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, a group of the foreground pixels constituting a foreground blob set of one or more foreground blobs; for each of N predetermined shapes at corresponding ones of N predetermined locations within the video image, where N is a positive integer greater than one, calculating a first value by comparing an overlap of the corresponding predetermined shape with the foreground blob set, the first value being used 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; and providing at least one of a report, an alarm, and an event detection when the crowd density exceeds a threshold value, using the determination of the representation of X humans, wherein a size of the corresponding predetermined shape for each of the N locations is determined in response to calibration of the video system, wherein the video system is used to obtain the video image, wherein the comparing of the corresponding predetermined shape with the foreground blob for each of the N predetermined shapes comprises analyzing an amount of an overlapping area of the corresponding predetermined shape with the foreground blob set. - View Dependent Claims (23, 24, 25, 26, 27, 28)
-
-
29. A video surveillance system, comprising:
-
a video source configured to provide a video image of a real world scene; and a computer, configured to detect foreground pixels of the video image, a group of the foreground pixels constituting a foreground blob set of one or more foreground blobs, wherein for each of N predetermined shapes at corresponding ones of N predetermined locations within the video image, where N is a positive integer greater than one, configured to calculate a first value by comparing an overlap of the corresponding predetermined shape with the foreground blob set, configured to determine X humans are represented by the foreground blob set using the first value, where X is whole number, configured to compare the corresponding predetermined shape with the foreground blob for each of the N predetermined shapes by analyzing an amount of an overlapping area of the corresponding predetermined shape with the foreground blob set and configured to provide at least one of a report, an alarm, and an event detection when the crowd density exceeds a threshold value, using the determined representation of X humans.
-
Specification