Apparatus for encoding a video signal using feature point based motion estimation

• US 5,751,362 A
• Filed: 10/17/1995
• Issued: 05/12/1998
• Est. Priority Date: 04/29/1995
• Status: Expired due to Fees

• Current Assignee: Daewoo Electronics Company Limited (Dayou Holdings Co.,Ltd.)
• Original Assignee: Daewoo Electronics Company Limited (Dayou Holdings Co.,Ltd.)
• Inventors: Lee, Min-Sup
• Primary Examiner(s): Chin, Tommy P.
• Assistant Examiner(s): Le, Vu
• Application Number:
  US08/544,355
• Time in Patent Office: 938 Days
• Field of Search: 348/699, 348/416, 348/413, 348/407, 348/402, 348/409, 348/412, 348/390, 348/415, 348/700
• US Class Current: 375/240.16
• CPC Class Codes: H04N 19/54 using feature points or meshes

Abstract

A motion-compensated video signal encoder has a circuit for determining a predicted current frame based on a current frame and a previous frame of a digital video signal. The circuit includes a region detection circuit for detecting a processing region encompassing a moving object from the previous frame based on a difference between the current and the previous frames to generate region information representing the detected processing region. Therefore, a number of pixels is selected from the pixels contained in the detected processing region as feature points based on the region information. A first set of motion vectors between the current and the previous frames, each of the first set of motion vectors representing a motion for each of the selected pixels is then detected. The first set of motion vectors is used for predicting the predicted current frame and transmitted as a set of motion vectors of the video signal together with the region information.

Claims

1. An apparatus, for use in a motion-compensated video signal encoder, for determining a predicted current frame based on a current frame and a previous frame of a digital video signal, comprising:

region detection means for calculating a difference between the current frame and the previous frame and for detecting a processing region encompassing a moving object from the previous frame based on the calculated difference to generate region information representing the detected processing region, wherein the processing region has a plurality of image blocks contained in the previous frame and the image blocks are determined by using the calculated difference, wherein said region detection means includes;

means for calculating the difference between the current and the previous frames on a pixel-by-pixel basis to generate a frame difference signal wherein the frame difference signal includes N×

M blocks, each block having P×

Q pixel difference values and N, M, P and Q are positive integers;

means for absolutizing the frame difference signal to generate an absolutized frame difference signal;

means for comparing the absolutized frame difference signal with a predetermined value to convert the absolutized frame difference signal into a converted frame difference signal, wherein, when a pixel difference value of the absolutized frame difference signal is smaller than the predetermined value, "0" is assigned as the pixel difference value and, otherwise, "1" is assigned as the pixel difference value;

means for counting the number of "1"'"'"'s contained in each block of the converted frame difference signal; and

means for comparing the counted number for each block with a predetermined number to generate the region information having sequentially arranged N×

M block representative values, wherein, when the counted number for a block is smaller than the predetermined value, "0" is assigned as a block representative value of the region information and, otherwise, "1" is assigned as the block representative value;

feature point selection means for selecting a number of pixels from the pixels contained in the detected processing region as feature points based on the region information;

first motion vector detection means for detecting a first set of motion vectors between the current and the previous frames, each of the first set of motion vectors representing a motion for each of the feature points;

second motion vector detection means for producing a second set of motion vectors for all of the pixels contained in the current frame by using the first set of motion vectors; and

motion compensation means for assigning the value of each of the pixels in the previous frame, said each of the pixels corresponding to one of the pixels in the current frame through one of the second set of motion vectors, as the value of said one of the pixels in the current frame, to thereby determine the predicted current frame.

2. The apparatus as recited in claim 1, wherein the second set of motion vectors includes a third set of motion vectors for all of the pixels contained in the processing region of the current frame, which is obtained based on the first set of motion vectors, and a fourth set of motion vector for the pixels lying outside the processing region of the current frame, which is made to consist of zero vectors.

3. The apparatus as recited in claim 2, wherein said feature point selecting means includes means for selecting a number of feature point pixels from the processing region of the previous frame by using a grid method which extracts pixels located at nodes of the grid as the feature point pixels.

4. The apparatus as recited in claim 3, wherein the grid is a rectangular grid.

5. The apparatus as recited in claim 4, wherein the grid is a hexagonal grid.

6. The apparatus as recited in claim 2, wherein said feature point selecting means includes means for selecting a number of feature point pixels from the processing region of the previous frame by using a grid and edge detection method which extracts pixels located at intersection points of the grid and the edge of the moving object as the feature point pixels.

7. The apparatus as recited in claim 6, wherein the third set of the motion vectors is obtained by using a known affine transform.

8. The apparatus as recited in claim 7, wherein said feature point selecting means includes means for selecting a number of feature point pixels from the processing region of the previous frame by using a grid and edge detection method which extracts pixels located at intersection points of the grid and the edge of the moving object as the feature point pixels.

9. The apparatus as recited in claim 8, wherein the third set of the motion vectors is obtained by using a known affine transform.

10. A video signal encoding apparatus for reducing a transmission rate of a digital video signal, said digital video signal having a plurality of frames including a current frame and a previous frame, comprising:

memory means for storing the previous frame of the digital video signal;

region detection means for calculating a difference between the current frame and the previous frame and for detecting a processing region encompassing a moving object from the previous frame based on the calculated difference to generate region information representing the detected processing region, wherein the processing region has a plurality of image blocks contained in the previous frame and the image blocks are determined by using the calculated difference, wherein said region detection means includes;

means for calculating the difference between the current and the previous frames on a pixel-by-pixel basis to generate a second frame difference signal wherein the second frame difference signal includes N×

M blocks, each block having P×

Q pixel difference values and N, M, P and Q are positive integers;

means for absolutizing the second frame difference signal to generate an absolutized frame difference signal;

means for comparing the absolutized frame difference signal with a predetermined value to convert the absolutized frame difference signal into a converted frame difference signal, wherein, when a pixel difference value of the absolutized frame difference signal is smaller than the predetermined value, "0" is assigned as the pixel difference value and, otherwise, "1" is assigned as the pixel difference value;

means for counting the number of "1"'"'"'s contained in each block of the converted frame difference signal; and

means for comparing the counted number for each block with a predetermined number to generate the region information having sequentially arranged N×

M block representative values, wherein, when the counted number for a block is smaller than the predetermined value, "0" is assigned as a block representative value of the region information and, otherwise, "1" is assigned as the block representative value;

feature point selection means for selecting a number of pixels from the pixels contained in the detected processing region as feature points based on the region information;

first motion vector detection means for detecting a first set of motion vectors between the current and the previous frames, each of the first set of motion vectors representing a motion for each of the selected pixels;

second motion vector detection means for producing a second set of motion vectors for all of the pixels contained in the current frame by using the first set of motion vectors;

motion compensation means for assigning the value of each of the pixels in the previous frame, said each of the pixels corresponding to one of the pixels in the current frame through one of the second set of motion vectors, as the value of said one of the pixels in the current frame, to thereby determine the predicted current frame; and

coding means for encoding a difference signal representing a difference between the current frame and the predicted current frame, the first set of motion vectors and the region information to thereby generate the digital video signal.

11. The apparatus as recited in claim 10, wherein the second set of motion vectors includes a third set of motion vectors for all of the pixels contained in the processing region of the current frame, which is obtained based on the first set of motion vectors, and a fourth set of motion vector for the pixels lying outside the processing region of the current frame, which is made to consist of zero vectors.

12. The apparatus as recited in claim 11, wherein the coding means includes:

means for generating a first frame difference signal by subtracting the predicted current frame from the current frame on a pixel-by-pixel basis;

means for encoding the first frame difference signal by using a discrete cosine transform and quantization circuit;

means for decoding the encoded differential pixel values to thereby provide a reconstructed frame difference signal;

means for providing a reconstructed current frame signal as the previous frame signal by combining the reconstructed frame difference signal and the predicted current frame signal; and

means for statistically coding the first frame difference signal, the first set of motion vectors and the region information.

13. The apparatus as recited in claim 12, wherein said feature point selecting means includes means for selecting a number of feature point pixels from the processing region of the previous frame by using a grid method which extracts pixels located at nodes of the grid as the feature point pixels.

14. The apparatus as recited in claim 13, wherein the grid is a rectangular grid.

15. The apparatus as recited in claim 14, wherein the grid is a hexagonal grid.