Cascading video object classification
First Claim
Patent Images
1. A method of constructing an object classifier in a camera system, comprising:
- receiving image data including representations of objects, the objects being labeled as either members or non-members of an object class in an object classifier, and generating metadata related to the objects;
processing the image data to extract a set of features for each of the objects;
for each of the features of the set of features for each of the objects, searching a set of discriminant functions to select for a given feature a discriminant function that optimizes separation of the labeled objects, the selected discriminant function being associated with its given feature to form one of a group of multiple feature/transformation combinations, wherein each of the discriminant functions comprises a weight vector and a feature vector; and
selecting for inclusion in the object classifier a feature/transformation combination from the group, the selected feature/transformation combination maximizing object classification performance of the object classifier in comparison to non-selected ones of the multiple feature/transformation combinations.
6 Assignments
0 Petitions
Accused Products
Abstract
A camera system comprises an image capturing device and an object classification module connected to the image capturing device. The object classification module is operable to determine whether or not an object in an image is a member of an object class. The object classification module includes multiple decision steps configured in a cascade configuration, wherein at least one of the multiple decision steps is operable to (a) accept an object as a member of the object class, (b) reject an object as a member of the object class, and (c) call on a next step to determine whether or not an object is a member of the object class.
-
Citations
8 Claims
-
1. A method of constructing an object classifier in a camera system, comprising:
-
receiving image data including representations of objects, the objects being labeled as either members or non-members of an object class in an object classifier, and generating metadata related to the objects; processing the image data to extract a set of features for each of the objects; for each of the features of the set of features for each of the objects, searching a set of discriminant functions to select for a given feature a discriminant function that optimizes separation of the labeled objects, the selected discriminant function being associated with its given feature to form one of a group of multiple feature/transformation combinations, wherein each of the discriminant functions comprises a weight vector and a feature vector; and selecting for inclusion in the object classifier a feature/transformation combination from the group, the selected feature/transformation combination maximizing object classification performance of the object classifier in comparison to non-selected ones of the multiple feature/transformation combinations. - View Dependent Claims (2, 3, 4)
-
-
5. A camera system, comprising:
-
a processor configured to receive image data including representations of objects, the objects being labeled as either members or non-members of an object class in an object classifier, to generate metadata related to the objects, and to process the image data to extract a set of features for each of the objects; for each of the features of the set of features for each of the objects, the object classifier being configured to search a set of discriminant functions to select for a given feature a discriminant function that optimizes separation of the labeled objects, the selected discriminant function being associated with its given feature to form one of a group of multiple feature/transformation combinations, wherein each of the discriminant functions comprises a weight vector and a feature vector, and the processor being further configured to select for inclusion in the object classifier a feature/transformation combination from the group, the selected feature/transformation combination maximizing object classification performance of the object classifier in comparison to non-selected ones of the multiple feature/transformation combinations. - View Dependent Claims (6, 7, 8)
-
Specification