Method and apparatus for encoding and decoding stereoscopic video

A method for decoding stereoscopic video including first and second images, comprising:

• (a) receiving an encoding stream, and outputting quantized data, of a video object of the first image, a motion vector, and quantized data of a horizontal disparity map;
• (b) decoding the video object and reconstructing the first image based on the quantized data of the video object and the motion vector;
• (c) decoding the quantized data of the horizontal disparity map based on the quantized data of the horizontal disparity map and the motion vector;
• and
• (d) performing disparity compensation based on the reconstructed first image and the decoded horizontal disparity map, and reconstructing the second image.

A method for encoding stereoscopic video including first and second images, comprising:

• (a) encoding the first image, and outputting a quantized video object and a motion vector of the first image;
• (b) decoding the quantized video object output in (a), and reconstructing the first image;
• (c) receiving the first and second images, and finding a pixel-based horizontal disparity map on the second image with reference to the first image;
• (d) encoding the horizontal disparity map and outputting a quantized horizontal disparity map based on the pixel-based horizontal disparity map and the motion vector;
• (e) reconstructing the quantized horizontal disparity map output in (d), and outputting a reconstructed horizontal disparity map;
• (f) performing disparity compensation and outputting a pixel value of a disparity-compensated second image based on a pixel value of the first image reconstructed in (b) and a horizontal disparity vector value of the horizontal disparity map reconstructed in (e);
• and
• (g) performing a residual process on the pixel value of the second image and the pixel value of the disparity-compensated second image output in (f) to output luminance residual texture, and encoding the luminance residual texture to output quantized luminance residual texture.

A method for decoding stereoscopic video including first and second images, comprising:

• (a) receiving an encoding stream, and outputting quantized data of a video object of the first image, a motion vector, quantized data of a horizontal disparity map, and quantized data of luminance residual texture;
• (b) decoding the video object and reconstructing the first image based on the quantized data of the video object and the motion vector;
• (c) decoding the quantized data of the horizontal disparity map based on the quantized data of the horizontal disparity map and the motion vector;
• (d) decoding the quantized data of the luminance residual texture based on the quantized data of the luminance residual texture and the motion vector;
• (e) performing disparity compensation based on the reconstructed first image and the decoded horizontal disparity map, and outputting disparity-compensated luminance texture;
• and
• (f) adding the disparity-compensated luminance texture and the luminance residual texture reconstructed in (d) to reconstruct the second image.

A method for encoding stereoscopic video including first and second images, comprising:

• (a) encoding the first image, and outputting a quantized video object and a motion vector of the first image;
• (b) decoding the quantized video object output in (a), and reconstructing the first image;
• (c) receiving the first and second images, and finding a pixel-based horizontal disparity map and a pixel-based vertical disparity map on the second image with reference to the first image;
• (d) encoding the horizontal disparity map and outputting a quantized horizontal disparity map based on the pixel-based horizontal disparity map and the motion vector;
• (e) encoding the vertical disparity map and outputting a quantized vertical disparity map based on the pixel-based vertical disparity map and the motion vector;
• (f) reconstructing the quantized horizontal disparity map output in (d), and outputting a reconstructed horizontal disparity map;
• (g) reconstructing the quantized vertical disparity map output in (d), and outputting a reconstructed vertical disparity map;
• (h) performing disparity compensation and outputting a pixel value of a disparity-compensated second image based on a pixel value of the first image reconstructed in (b), a horizontal disparity vector value of the horizontal disparity map reconstructed in (f), and a vertical disparity vector value of the vertical disparity map reconstructed in (h);
• and
• (i) performing a residual process on the pixel value of the second image and the pixel value of the disparity-compensated second image output in (h) to output luminance residual texture, and encoding the luminance residual texture to output quantized luminance residual texture.

A method for decoding stereoscopic video including first and second images, comprising:

• (a) receiving an encoding stream, and outputting quantized data of a video object of the first image, a motion vector, quantized data of a horizontal disparity map, quantized data of a vertical disparity map, and quantized data of luminance residual texture;
• (b) decoding the video object and reconstructing the first image based on the quantized data of the video object and the motion vector;
• (c) decoding the quantized data of the horizontal disparity map based on the quantized data of the horizontal disparity map and the motion vector;
• (d) decoding the quantized data of the vertical disparity map based on the quantized data of the vertical disparity map and the motion vector;
• (e) decoding the quantized data of the luminance residual texture based on the quantized data of the luminance residual texture and the motion vector;
• (f) performing disparity compensation based on the reconstructed first image, the decoded horizontal disparity map, and the decoded vertical disparity map, and outputting disparity-compensated luminance texture;
• and
• (g) adding the disparity-compensated luminance texture and the luminance residual texture reconstructed in (e) to reconstruct the second image.

An encoder for stereoscopic video including first and second images, comprising:

• a video object encoder for encoding the first image, and outputting a quantized video object and a motion vector of the first image;
• a disparity estimator for receiving the first and second images, and finding a pixel-based horizontal disparity map on the second image with reference to the first image;
• and
• an auxiliary component encoder for encoding the horizontal disparity map and outputting a quantized horizontal disparity map based on the pixel-based horizontal disparity map output by the, disparity estimator and a motion vector output by the video object encoder.

A decoder for stereoscopic video including first and second images, comprising:

• a variable length decoder for receiving an encoding stream, and outputting quantized data of a video object of the first image, a motion vector, and quantized data of a horizontal disparity map;
• a video object decoder for decoding the video object and reconstructing the first image based on the quantized data of the video object and the motion vector output by the variable length decoder;
• an auxiliary component decoder for decoding the quantized data of the horizontal disparity map based on the quantized data of the horizontal disparity map and the motion vector output by the variable length decoder;
• and
• a disparity compensator for performing disparity compensation based on the reconstructed first image output by the video object decoder and the decoded horizontal disparity map output by the auxiliary component decoder, and reconstructing the second image.

An encoder for stereoscopic video including first and second images, comprising:

• a video object encoder for encoding the first image to output a quantized video object and a motion vector of the first image, and encoding the quantized video object to output a reconstructed first image;
• a disparity estimator for receiving the first and second images, and finding a pixel-based horizontal disparity map on the second image with reference to the first image;
• a first auxiliary component encoder for encoding the horizontal disparity map to output a quantized horizontal disparity map, and decoding the output and quantized horizontal disparity map to output a reconstructed horizontal disparity map based on the pixel-based horizontal disparity map output by the disparity estimator and the motion vector output by the video object encoder;
• a disparity compensator for performing disparity compensation and outputting a pixel value of a disparity-compensated second image based on a pixel value of the reconstructed first image output by the video object encoder and a horizontal disparity vector value of the, reconstructed horizontal disparity map output by the first auxiliary component encoder;
• and
• a second auxiliary component encoder for performing a residual process on the pixel value of the second image and the pixel value of the disparity-compensated second image output by the disparity compensator to output luminance residual texture, and encoding the luminance residual texture to output quantized luminance residual texture.

A decoder for stereoscopic video including first and second images, comprising:

• a variable length decoder for receiving an encoding stream, and outputting quantized data of a video object of the first image, a motion vector, quantized data of a horizontal disparity map, and quantized data of luminance residual texture;
• a video object decoder for decoding the video object and reconstructing the first image based on the quantized data of the video object and the motion vector;
• a first auxiliary component decoder for decoding the quantized data of the horizontal disparity map based on the quantized data of the horizontal disparity map and the motion vector;
• a second auxiliary component decoder for decoding the quantized data of the luminance residual texture based on the quantized data of the luminance residual texture and the motion vector;
• a disparity compensator for performing disparity compensation based on the reconstructed first image output by the video object decoder and the decoded horizontal disparity map output by the first auxiliary component decoder, and outputting disparity-compensated luminance texture and disparity-compensated chrominance texture;
• and
• a first adder for adding the disparity-compensated luminance texture output by the disparity compensator and the reconstructed luminance residual texture output by the second auxiliary component decoder.