Video seed fill over time

• US 5,706,361 A
• Filed: 01/26/1995
• Issued: 01/06/1998
• Est. Priority Date: 01/26/1995
• Status: Expired due to Fees

• Current Assignee: Autodesk, Inc.
• Original Assignee: Autodesk, Inc.
• Inventors: de Freitas Garcia, Jose Madeira, Kent, William James, Jellinghaus, Robert Norman
• Primary Examiner(s): COUSO, YON JUNG
• Application Number:
  US08/378,663
• Time in Patent Office: 1,076 Days
• Field of Search: 382/173, 382/164, 382/282, 382/283, 382/103, 382/308, 382/291, 382/294, 395/152, 348/34
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/187 involving region growing; involving region merging; involving connected component labelling

Abstract

A method for identifying the limits of an object (10) in a digitized video display (12) having a series of video frames (14). A seed (26) is introduced into an initial object (18a) in an initial frame (14b) and the seed (26) is expanded to define the limits of the initial object (18a). The modified seed (26a) is then superimposed upon a modified object (18b) and a common area (36) is identified. The common area (36) is modified to become yet another iteration of the seed (26) for continuing the operation into yet another frame (14) having a subsequent version of the object (18). The video seed fill over time method (10) proceeds until no common area (36) is identified for a specified number of frames 14.

Claims

1. A method for identifying an object within a plurality of frames of a moving picture video, comprising:

identifying the object within a first frame of the video according to a distinguishing characteristic of the object and further identifying a location of the object;

superimposing the location of the object within a second frame of the video;

identifying a common area within the second frame of the video;

expanding the common area to identify the location of a modified object within the second frame of the video;

superimposing the location of the modified object within a third frame of the video;

identifying a second common area within the third flame of the video; and

expanding the common area to identify a second modified object within the third frame of the video.

2. The method of claim 1, wherein:

the distinguishing characteristic is color.

3. The method of claim 1, wherein:

the distinguishing characteristic is brightness.

4. The method of claim 1, wherein:

the distinguishing characteristic is texture.

5. The method of claim 1, wherein:

said common area is that portion of the second frame having both the distinguishing characteristic and the location of the object.

6. A method for isolating an object through a series of video frames, comprising:

a choosing an identifying criterion;

b placing a seed within the object in an initial frame;

c1 expanding said seed to define the limits of the object according to the identifying criterion;

c2 saving the location of the seed;

d advancing to a next frame;

e superimposing the seed on said next frame;

f identifying a common area having both the seed and the identifying criterion; and

g modifying the seed to coincide with the common area.

7. The method of claim 6, and further including:

repeating steps c1 through g until no common area can be identified in step f.

8. The method of claim 6, and further including:

repeating steps c1 through g until no common area can be identified in step f, and then repeating steps d through f a predetermined number of times or else until a common area is identified in step f.

9. The method of claim 6, wherein:

the next frame is immediately subsequent to the initial frame in the series of video frames.

10. The method of claim 6, wherein:

the next frame is subsequent to the initial frame in the series of video frames.

11. The method of claim 6, wherein:

the next frame is previous to the initial frame in the series of video frames.

12. An object defining system for defining a portion of a moving picture video image, comprising:

display means for displaying a first frame of a computer image;

selection means for an operator to place a seed upon the first frame; and

a computer program including operations for;

a expanding the seed to define an object;

b redefining the seed to correspond to the location of the object and saving a location of the seed;

c advancing to a next frame;

d overlaying the location of the seed on the next frame;

e identifying a common area, the common area being any portions of the next frame which are shared by the object and the seed; and

f redefining the seed to correspond to the common area.

13. The object defining system of claim 12, wherein:

the computer program is caused to repeat steps a through f until no common area is identified in step e.

14. The object defining system of claim 12, wherein:

the computer program is caused to repeat steps a through f even after no common area is identified in step e.

15. The object defining system of claim 12, wherein:

the seed is defined according to a characteristic of that portion of the first frame upon which the operator places the seed.

16. The object defining system of claim 15, wherein:

the characteristic of that portion of the first frame upon which the operator places the seed is the color of that portion of the first frame.

17. The object defining system of claim 15, wherein:

the seed is defined according to a single color.

18. The object defining system of claim 15, wherein:

the seed is initially defined according to a characteristic of the object.