Image object ranking

• US 20020057343A1
• Filed: 06/29/2001
• Published: 05/16/2002
• Est. Priority Date: 06/30/2000
• Status: Active Grant

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Finchbaugh, Bruce E., Ronk, Lawrence J.
• Granted Patent: US 7,386,170 B2
• Time in Patent Office: Days
• US Class Current: 348/169
• CPC Class Codes: G06F 16/5838 using colour
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 7/223 using block-matching

Abstract

Automatic vision system object indexing and image database query system using both path-dependent and path-independent features of moving objects within a sequence of images. Feature vectors of both average over frames of an object traversing the field of view plus average over blocks of a grid for a path association. Color histograms may be an included feature.

Claims

1. A method of video object feature data generation, comprising:

(a) extracting a first set of features from a moving object detected in a sequence of images;

(b) extracting a sequence of grid blocks corresponding to motion of said object in said sequence of images; and

(c) storing said first set of features and said sequence of grid blocks.

2. The method of claim 1, wherein:

(a) said extracting of step (a) of claim 1 includes extracting features in every image in said sequence containing said object.

3. The method of claim 1, further comprising:

(a) for each of said grid blocks of step (b) of claim 1, extracting features and associating said grid-block extracted features with said grid block sequence.

4. The method of claim 1, wherein:

(a) said first set of extracted features of step (a) of claim 1 includes a color histogram.

5. A method of searching for a video object, comprising:

(a) providing a database of feature vectors of video objects;

(b) providing a target feature vector;

(c) comparing said target feature vector to each feature vector of said database;

(d) ranking said feature vectors of said database according to the results of step (c); and

(e) finding video objects by an association of video objects with said feature vectors of said database together with the results of step (d).

6. The method of claim 5, wherein:

(a) said feature vectors of said database include both path-dependent and path-independent features extracted from video objects moving in sequences of images.

7. The method of claim 6, wherein:

(a) said path-dependent features for a video object include a sequence of grid blocks traversed by said video object together with averages of features extracted from said video object when said video object was located in a corresponding grid block.