[METHOD OF MOTION DETECTION FOR 3D COMB FILTER VIDEO DECODER]

• US 20050134744A1
• Filed: 03/30/2004
• Published: 06/23/2005
• Est. Priority Date: 12/22/2003
• Status: Active Grant

• Current Assignee: Sunplus Technology Co., Ltd
• Original Assignee: Sunplus Technology Co., Ltd
• Inventors: PENG, URIAH, SHAN, PEI-MING
• Granted Patent: US 7,432,987 B2
• Time in Patent Office: Days
• US Class Current: 348/670
• CPC Class Codes: H04N 9/78 for separating the brightness signal or the chrominance signal from the colour television signal, e.g. using comb filter

Abstract

A method of motion detection for a 3D comb filter video decoder is disclosed. In this method, a plurality of sampled data F_mP_x,y is obtained and temporarily stored after a composite video signal is sampled, wherein F_mP_x,y represents a sampled data of the y^th pixel on the x^th line of the m^th frame inside the composite video signal, and m, x, y are positive integers greater than or equal to 0. Then, F_m+1P_x,y, F_mP_x,y, F_m−1P_x,y, and F_m−2P_x,y are used to determine a motion/still status of the composite video signal. Since the present invention performs the motion detection according to the composite video signal whose Y/C has not been separated yet, the present invention can accurately determine the motion level.

Claims

1. A method of motion detection for a 3D comb filter video decoder, comprising:

sampling a composite video signal for obtaining a plurality of temprarily stored sampled data F_mP_x,y, wherein F_mP_x,y represents a sampled data of a y^th pixel on an x^th line of an m^th frame in the composite video signal, and m, x, y are positive integers greater than or equal to 0; and

using F_m+1P_x,y, F_mP_x,y, F<sub>m−

1</sub>P_x,y, and F<sub>m−

2</sub>P_x,y to determine a motion/still status of the composite video signal.

2. The method of motion detection for a 3D comb filter video decoder of claim 1, wherein the step of determining the motion/still status of the composite video signal further comprises:

using F_m+1P_x,y, F_mP_x,y, F<sub>m−

1</sub>P_x,y, and F<sub>m−

2</sub>P_x,y to calculate and obtain a plurality of max differences MD_x,y, wherein MD_x,y represents a max difference of the y^th pixel on the x^th line;

averaging 4 max differences of the contiguous pixels selected to obtain a motion factor MF_x,y, wherein MF_x,y represents a motion factor of the y^th pixel on the x^th line; and

detecting MF_x,y to determine the motion/still status of the y^th pixel on the x^th line in the composite video signal.

3. The method of motion detection for a 3D comb filter video decoder of claim 2, wherein when it is determined that the composite video signal is a signal for an NTSC system, the step of sampling the composite video signal uses a frequency which is 4 times the subcarrier frequency in the composite video signal to sample the signal, and the signal is sampled when the subcarrier phase is equal to 0, 0.5π

, π

, and 1.5π

.

4. The method of motion detection for a 3D comb filter video decoder of claim 3, wherein MD_x,y is calculated based on an equation:

MD_x,y=Max{|F_mP_x,y−

F<sub>m−

2</sub>P_x,y|, |F_m+1P_x,y−

F<sub>m−

1</sub>P_x,y|}.

5. The method of motion detection for a 3D comb filter video decoder of claim 2, wherein when it is determined that the composite video signal is a signal for a PAL system, the step of sampling the composite video signal uses a frequency which is 4 times the subcarrier frequency in the composite video signal to sample the signal, and the signal is sampled when the subcarrier phase is equal to 0.25π

, 0.75π

, 1.25π

, and 1.75π

.

6. The method of motion detection for a 3D comb filter video decoder of claim 5, wherein the step of calculating and obtaining MD_x,y further comprises:

calculating and obtaining a plurality of luma differences LD_x,y, wherein LD_x,y represents a luma difference of the y^th pixel on the x^th line, and is calculated based on an equation;

LD_x,y=|F_mP_x,y+F<sub>m−

2</sub>P_x,y-F_m+1P_x,y−

F<sub>m−

1</sub>P_x,y|;

calculating and obtaining a plurality of intermediate differences IMD_x,y, wherein IMD_x,y represents an intermediate difference of the y^th pixel on the x^th line, and is calculated based on an equation;

IMD<sub>i,2j−

1</sub>=Max{|F_m+1P<sub>i,2j−

1</sub>−

F<sub>m−

2</sub>P<sub>i,2j−

1</sub>|, |F_mP<sub>i,2j−

1</sub>−

<sub>m−

1</sub>P<sub>i,2j−

1</sub>|};


IMD_i,2j=Max{|F_m+1P_i,2j−

FP_i,2j|, |F<sub>m−

1</sub>P_i,2jF<sub>m−

2</sub>P_i,2j|}; and

calculating and obtaining MD_x,y, which is calculated based on an equation;

MD_x,y=a*IMD+(1−

a)*LD_x,y;

wherein, a is a real number greater than 0 and less than 1, and i, j are positive integers.

7. The method of motion detection for a 3D comb filter video decoder of claim 2, wherein the step of obtaining MF_x,y further comprises:

averaging 4 max differences of the contiguous pixels selected to obtain a plurality of max differences AMD_x,h, wherein AMD_x,h represents an average of max difference of a h^th pixel on the x^th line, h is a positive integer, and AMD_x,h is calculated based on an equation;

AMD_x,h=(MD_x,h+MD_x,h+1+MD_x,h+2+MD_x,h+3)/4;

and selecting a minimum from the averages of max difference to obtain a motion factor MF_x,y, wherein MF_x,y represents a motion factor of the y^th pixel on the x^th line.

8. The method of motion detection for a 3D comb filter video decoder of claim 7, wherein the step of selecting a minimum from the averages of max difference to obtain MF_x,y is based on an equation:

MF_x,y=Min(AMD_x,y, AMD<sub>x,y−

1</sub>, AMD<sub>x,y−

2</sub>, AMD<sub>x,y−

3</sub>).

9. The method of motion detection for a 3D comb filter video decoder of claim 7, wherein the step of selecting a minimum from the averages of max difference to obtain MF_x,y is based on an equation:

MF_x,y=Min(AMD_x,y, AMD<sub>x,y−

3</sub>).

10. The method of motion detection for a 3D comb filter video decoder of claim 2, wherein the step of detecting MF_x,y to determine the motion/still status of the y^th pixel on the x^th line in the composite video signal further comprises:

providing a threshold; and

comparing MF_x,y with the threshold, and when MF_x,y is greater than the threshold, it is determined that the y^th pixel on the x^th line in the composite video signal is in the motion status, otherwise, the y^th pixel on the x^th line in the composite video signal is in the still status.

11. The method of motion detection for a 3D comb filter video decoder of claim 10, wherein the motion factors MF_x,y are the motion factors of the m^th frame.