Method and apparatus for reducing the bit rate in a video object planes sequence coder

Claim 3

The method of claim 1, wherein the width of said outline stripe has a maximum value corresponding to a maximum distance between the picture elements on a boundary of the predicted image and corresponding picture elements contained on the boundary of said approximate shape.

Claim 4

The method of claim 2, wherein the width of said outline strip is a constant value.

Claim 5

The method of claim 2, wherein the width of said outline stripe varies based on the boundary and size of said approximate shape of the object in the current frame and the prior frame and the displacements of the object.

Claim 6

The method of claim 3, wherein the width value is below a threshold width value.

Claim 7

The method of claim 4, wherein the maximum width value is below a threshold width value.

Claim 8

The method of claim 1, wherein said assigning step is performed using picture element repetition techniques.

Claim 9

The method of claim 3, wherein said assigning step is performed using picture element repetition techniques.

Claim 10

The method of claim 4, wherein said assigning step is performed using picture element repetition techniques.

Claim 11

The method of claim 1, wherein said assigning step is performed utilizing texture extrapolation techniques.

Claim 12

The method of claim 3, wherein said assigning step is performed utilizing texture extrapolation techniques.

Claim 13

The method of claim 4, wherein said assigning step is performed utilizing texture extrapolation techniques.

Claim 14

The method of claim 1, further comprising the step of dividing each video frame into macro blocks, and wherein said predicting step further comprises utilizing block motion compensation techniques for predicting, for a particular block containing at least a portion of the image in the current frame, texture of the image portion in the particular block based on (1) displacement of the image-portion containing macro block from the previous frame to the current frame, (2) the predicted image of the image-portion containing macro block from the previous frame and (3) the masked error signal for the image-portion containing macro block of the prior frame.

Claim 15

The method of claim 9, further comprising the step of dividing each video frame into macro blocks, and wherein said predicting step further comprises utilizing block motion compensation techniques for predicting, for a particular block containing at least a portion of the image in the current frame, texture of the image portion in the particular block based on (1) displacement of the image-portion containing macro block from the previous frame to the current frame, (2) the predicted image of the image-portion containing macro block from the previous frame and (3) the masked error signal for the image-portion containing macro block of the prior frame.

Claim 16

The method of claim 10, further comprising the step of dividing each video frame into macro blocks, and wherein said predicting step further comprises utilizing block motion compensation techniques for predicting, for a particular block containing at least a portion of the image in the current frame, texture of the image portion in the particular block based on (1) displacement of the image-portion containing macro block from the previous frame to the current frame, (2) the predicted image of the image-portion containing macro block from the previous frame and (3) the masked error signal for the image-portion containing macro block of the prior frame.

Claim 17

The method of claim 1, further comprising the step of dividing each video frame into a plurality of macro blocks, and wherein said predicting step further comprises utilizing overlapped block motion compensation techniques for predicting, for a particular block containing at least a portion of the object image in the current frame, texture of the object image portion in the particular block.

Claim 18

The method of claim 1, further comprising the step of dividing each video frame into a plurality of macro blocks, and wherein said predicting step further comprises utilizing overlapped block motion compensation techniques for predicting, for a particular block containing at least a portion of the object image in the current frame, texture of the object image portion in the particular block based on (1) displacement of the image-portion containing macro block from the previous frame to the current frame, (2) the predicted image of the image-portion containing macro block from the previous frame, (3) predicted images of portions of macro blocks in the previous frame which are adjacent the image-portion containing macro block of the previous frame, (4) the masked error signal for the image-portion containing macro block of the prior frame, (5) the masked error signals of macro blocks adjacent the image-portion containing macro block of the previous frame, and (6) the estimated displacements of said adjacent blocks of the current frame.

Claim 19

The method of claim 17, wherein the portions of macro blocks in the previous frame which are adjacent the image portion containing macro block of the previous frame, contain sections of the object image.

Claim 20

The method of claim 10, further comprising the step of dividing each video frame into a plurality of macro blocks, and wherein said predicting step further comprises utilizing overlapped block motion compensation techniques for predicting, for a particular block containing at least a portion of the image in the current frame, texture of the image portion in the particular block.

Claim 21

The method of claim 10, further comprising the step of dividing each video frame into a plurality of macro blocks, and wherein said predicting step further comprises utilizing overlapped block motion compensation techniques for predicting, for a particular block containing at least a portion of the image in the current frame, texture of the image portion in the particular block based on (1) displacement of the image-portion containing macro block from the previous frame to the current frame, (2) the predicted image of the image-portion containing macro block from the previous frame, (3) predicted images of portions of macro blocks in the previous frame which are adjacent the image-portion containing macro block of the previous frame, (4) the masked error signal for the image-portion containing macro block of the prior frame, (5) the masked error signals of macro blocks adjacent the image-portion containing macro block of the previous frame, and (6) the estimated displacements of said adjacent blocks of the current frame.

Claim 22

The method of claim 20, wherein the portions of macro blocks in the previous frame which are adjacent the image portion containing macro block of the previous frame, contain sections of the object image.

Claim 23

The method of claim 1, further comprising the step of masking said predicted image and said outline stripe with said shape approximation to form a resulting prediction signal, and storing said prediction signal in memory along with said masked error signal for use in predicting an image for the object in a subsequent frame.

Claim 24

The method of claim 1, further comprising the step of eroding the shape approximation to remove therefrom any non-object picture elements, said eroding step being performed prior to said predicting step.

Claim 25

A video object planes sequence encoder, having a reduced output bit rate, for analyzing and coding an object image having texture and position components and represented by a video signal defined by a continuous stream of successive-in-time video frames having a plurality of picture elements, said encoder comprising: a shape approximator for approximating a shape of an object as it appears on a current frame to define an approximate shape having a boundary and a size; a displacement estimator for estimating frame-to-frame displacements of the object based on object position in a prior frame and object position in the current frame for generating object displacement estimates; an image predictor connected to said shape approximator and said displacement estimator for predicting an image for the object in the current frame based on an image of the object on said prior frame, said approximate shape and said object displacement estimates; an outline generator for generating an outline stripe having a width and containing picture elements, said stripe encompassing said predicted image so that said predicted image and said stripe are sized at least equal to said approximate shape; means for assigning texture values to the picture elements in said stripe, said values being based on picture element values in said approximate shape; means for subtracting said outline stripe and said predicted image from the current frame to form an error signal; a mask unit connected to said subtracting means for masking said error signal with said approximate shape to pass substantially all picture elements that are active in both said error signal and said approximate shape; a quantizer in communication with said mask unit for quantizing the masked error signal; and an entropy coder for coding the quantized masked error signal for transmission at an output bit rate.

Claim 26

A video object planes sequence encoder, having a reduced output bit rate, for analyzing and coding an object image having texture and position components and represented by a video signal defined by a continuous stream of successive-in-time video frames having a plurality of picture elements, said encoder comprising: a shape approximator for approximating a shape of an object as it appears on a current frame to define an approximate shape having a boundary and a size; a displacement estimator for estimating frame-to-frame displacements of the object based on object position in a prior frame and object position in the current frame for generating object displacement estimates; an image predictor connected to said shape approximator and said displacement estimator for predicting an image for the object in the current frame based on an image of the object on said prior frame, said approximate shape and said object displacement estimates; an outline generator for generating an outline stripe having a width and containing picture elements, said stripe encompassing said predicted image so that said predicted image and said stripe are sized at least equal to said approximate shape; means for assigning texture values to the picture elements in said stripe, said values being based on picture element values in said approximate shape; means for subtracting said outline stripe and said predicted image from the current frame to form an error signal; a mask unit connected to said subtracting means for masking said error signal with said approximate shape to pass substantially all picture elements that are active in both said error signal and said approximate shape; a quantizer in communication with said mask unit for quantizing the masked error signal; and an entropy coder for coding the quantized masked error signal for transmission at an output bit rate.

Claim 27

The encoder of claim 24, wherein the width of said outline stripe has a maximum value corresponding to a maximum distance between the picture elements on a boundary of the predicted image and corresponding picture elements contained on the boundary of said approximate shape.

Claim 28

The encoder of claim 25, wherein the maximum width value is below a threshold width value.

Claim 29

The encoder of claim 24, further comprising a second mask unit connected to said shape approximator and said image predictor for forming a predictor signal by masking said predicted image and stripe with said approximate shape to pass substantially all picture elements that are active in both said predicted image and stripe and said approximate shape, and a frame storage means connected to said second mask unit for storing said predictor signal for use in predicting an image of the object in a subsequent frame.

Claim 30

The encoder of claim 27, further comprising an erosion device connected between said shape approximator and said image predictor for eroding the approximate shape prior to input to said image predictor.