METHOD FOR DETECTING A MOVING OBJECT IN MOTION VIDEO AND APPARATUS THEREFOR

• US 20080159406A1
• Filed: 02/25/2008
• Published: 07/03/2008
• Est. Priority Date: 09/02/1999
• Status: Abandoned Application

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: KOBAYASHI, Hiroyuki, CHUJOH, Takeshi, KIKUCHI, Yoshihiro, SAKUMA, Akira, HAYASHI, Toshifumi
• Application Number:
  US12/036,966
• Publication Number:
  US 20080159406A1
• Time in Patent Office: Days
• US Class Current: 375/240.24
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 7/254 involving subtraction of images

Abstract

An apparatus for detecting a moving object in motion video comprises a macro-block determining section for determining the background/non-background of each macro-block of a reconstructed video signal from a video decoder section which decodes encoded data obtained by compression-encoding a motion video signal, a moving object determining section for determining an area of the moving object from the result of the determination on the background/non-background, and a moving object combination display for displaying information indicating the area of the moving object on a display screen for the reconstructed video signal. The macro-block determining section determines if a macro-block represents a background area or a non-background area, based on mode information from the video decoder section and a cross correlation value between a present frame of the reconstructed video signal and a signal of a frame preceding the present frame by one frame, obtained by a first cross correlation calculator, and a cross correlation value between the present frame of the reconstructed video signal and a background video signal stored in a background memory, obtained by a second cross correlation calculator.

Claims

1. A method for detecting a video moving object comprising:

(a) determining whether a video signal corresponding to a unit area represents a background area or a non-background area, the video signal being part of a reconstructed video signal acquired by a video decoder which decodes encoded data obtained by compression-encoding a motion video signal;

(b) detecting adjacent unit areas representing the non-background area using a determination result of the determining for each unit area to determine the adjacent unit areas as an area of a moving object;

(c) computing a cross correlation value between a current frame of the reconstructed video signal and a video signal preceding by one frame every unit area,

(d) storing, a storage unit, a background video signal corresponding to a background component of the reconstructed video, and

(e) computing a second cross correlation value between the current frame of the reconstructed video signal and the background video signal stored in the storage unit every unit area,

wherein the determining (a) includes (f) determining whether the video signal in the unit area is a background area or a nonbackground area on the basis of encoding mode information obtained from the video decoder, the first cross correlation value and the second cross correlation value.

2. The method according to claim 1, wherein the determining (a) includes (g) determining whether an interest macro block corresponding to the unit area is a background macro block or a non-background macro block every frame, and the detecting (b) includes detecting the moving object on the basis of a determination result as the background area;

wherein the determining (g) includes,

computing a first cross correlation value between a local decoded picture signal and a reference picture signal of a frame preceding by one frame,

computing a second cross correlation value between the local decoded picture signal and a background picture signal preceding by one frame, and

determining a background or a non-background every macro block in the frame on the basis of decoded mode information, the first cross correlation value, and the second cross correlation value.

3. The method according to claim 2, wherein the determining (g) includes determining the interest macro block as a background macro block when the first cross correlation value is larger than a first threshold.

4. The method according to claim 2, wherein the determining (g) includes determining the interest macro block as a non-background macro block when the second cross correlation value is larger than a second threshold, and as a background macro block when the second cross correlation value is not more than the second threshold.