Method and apparatus for encoding image and method and apparatus for decoding image using human visual characteristics
AssigneeSamsung Electronics Co., Ltd.
InventorWoo-jin Han
Life of this patent
Estimating the term…
Abstract
An image encoding method and apparatus capable of improving image compression efficiency by using human psycho-visual characteristics are provided. In the image encoding method and apparatus, an image signal component of which amplitude is equal to or less than a just noticeable distortion (JND) that is a minimum limit of a visually noticeable image is not encoded according to image characteristics by using the masking effect of the human visual system (HVS), thereby improving an image compression rate without visible deterioration of image quality.
Claims35
An image encoding method comprising:
- creating a predicted block corresponding to a current block to be encoded using a prediction unit;
- creating a residual block by estimating a difference between the current block and the predicted block using a subtraction unit;
- creating a transformed residual block and a transformed predicted block by transforming the residual block and the predicted block, respectively using a transform unit;
- using a minimum threshold block creating unit for obtaining a minimum threshold transform coefficient which allows a corresponding pixel included in the transformed predicted block to be visually noticeable, and creating a minimum threshold block comprising the minimum threshold transform coefficient;
- and
- using a quantization unit for determining an optimum quantization step which allows a maximum quantization error to be equal to or less than the minimum threshold transform coefficient with respect to a corresponding coefficient of the transformed residual block, and quantizing the corresponding coefficient of the transformed residual block based on the determined optimum quantization step.
An image encoding method comprising:
- creating a predicted block corresponding to a current block to be encoded using a prediction unit;
- creating a residual block by estimating a difference between the current block and the predicted block using a subtraction unit;
- creating a transformed residual block and a transformed predicted block by transforming the residual block and the predicted block, respectively using a transform unit;
- using a minimum threshold block creating unit for obtaining a minimum threshold transform coefficient which allows a corresponding pixel of the transformed predicted block to be visually noticeable, and creating a minimum threshold block comprising the minimum threshold transform coefficient;
- and
- using a selection unit for selecting a coefficient of the transformed residual block to be encoded from coefficients of the transformed residual block by comparing the minimum threshold transform coefficient with a corresponding coefficient of the transformed residual block.
An image encoding method comprising:
- creating a predicted block corresponding to a current block to be encoded using a prediction unit;
- creating a residual block by estimating a difference between the current block and the predicted block using a first subtraction unit;
- using a minimum threshold block creating unit for obtaining a minimum threshold coefficient which allows a corresponding pixel included in the predicted block to be visually noticeable, and creating a minimum threshold block comprising the minimum threshold coefficient;
- and
- using a second subtraction unit for creating a difference residual block by estimating a difference between the minimum threshold coefficient and a corresponding coefficient of the residual block.
An image encoding apparatus comprising:
- a prediction unit which creates a predicted block corresponding to a current block to be encoded;
- a subtraction unit which creates a residual block by estimating a difference between the current block and the predicted block;
- a first transform unit which creates a transformed residual block by transforming the residual block;
- a second transform unit which creates a transformed predicted block by transforming the predicted block;
- a minimum threshold block creating unit which obtains a minimum threshold transform coefficient which allows a coefficient included in the transformed predicted block to be visually noticeable, and creates a minimum threshold block comprising the minimum threshold transform coefficient;
- and
- a quantization unit which determines an optimum quantization step which allows a maximum quantization error to be equal to or less than the minimum threshold transform coefficient with respect to a corresponding coefficient of the transformed residual block, and quantizing the corresponding coefficient of the transformed residual block based on the determined optimum quantization step.
An image encoding apparatus comprising:
- a prediction unit which creates a predicted block corresponding to a current block to be encoded;
- a subtraction unit which creates a residual block by estimating a difference between the current block and the predicted block;
- a first transform unit which creates a transformed residual block by transforming the residual block;
- a second transform unit which creates a transformed predicted block by transforming the predicted block;
- a minimum threshold block creating unit which obtains a minimum threshold transform coefficient which allows a coefficient included in the transformed predicted block to be visually noticeable, and creates a minimum threshold block comprising the minimum threshold transform coefficient;
- and
- a selecting unit selecting a coefficient of a transformed residual block to be encoded from coefficients of the transformed residual block by comparing the minimum threshold transform coefficient with a corresponding coefficient of the transformed residual block.
An image encoding apparatus comprising:
- a prediction unit which creates a predicted block corresponding to a current block to be encoded;
- a first subtraction unit which creates a residual block by estimating a difference between the current block and the predicted block;
- a minimum threshold block creating unit which obtains a minimum threshold coefficient which allows a corresponding pixel included in the predicted block to be visually noticeable, and creates a minimum threshold block comprising the minimum threshold coefficient;
- and
- a second subtraction unit which creates a difference residual block by estimating a difference between the minimum threshold coefficient and a corresponding coefficient of the residual block.
An image decoding method comprising:
- using an entropy decoder for reading information on an optimum quantization step, which is determined such that a maximum quantization error is equal to or less than a minimum threshold transform coefficient with respect to each coefficient of a transformed residual block, and information on a coefficient of the transformed residual block, from a received bit-stream, the coefficient being quantized according to the optimum quantization step;
- inverse-quantizing the quantized coefficient of the transformed residual block using an inverse quantization unit;
- and
- restoring an image by adding a predicted block for a current block to be decoded and the inverse-transformed transformed residual block using an add unit.
An image decoding apparatus comprising:
- an entropy decoder which reads information on an optimum quantization step, which is determined such that a maximum quantization error is equal to or less than a minimum threshold transform coefficient with respect to each coefficient of a transformed residual block, and information on a coefficient of the transformed residual block, from a received bit-stream, the coefficient being quantized according to the optimum quantization step;
- an inverse-quantization unit inverse-quantizing the coefficients of the transformed residual block, which are quantized according to the read optimum quantization step;
- an inverse-transform unit which outputs a residual block by inverse-transforming the quantized coefficient of the transformed residual block;
- a prediction unit which creates a predicted block for a current block to be decoded;
- and
- an add unit which restores an image by adding the inverse-transformed residual block and the predicted block.
An image encoding method comprising:
- creating a predicted image corresponding to a current image to be encoded using a prediction unit;
- creating a residual image by estimating a difference between the current image and the predicted image using a subtraction unit;
- creating a transformed residual image and a transformed predicted image by transforming the residual image and the predicted image, respectively using a transform unit;
- using a minimum threshold block creating unit for obtaining a minimum threshold transform coefficient which allows a corresponding pixel included in the transformed predicted image to be visually noticeable, and creating a minimum threshold image comprising the minimum threshold transform coefficient;
- and
- using a quantization unit for determining an optimum quantization step which allows a maximum quantization error to be equal to or less than the minimum threshold transform coefficient with respect to a corresponding coefficient of the transformed residual image, and quantizing the corresponding coefficient of the transformed residual image based on the determined optimum quantization step.
An image encoding method comprising:
- creating a predicted image corresponding to a current image to be encoded;
- creating a residual image by estimating a difference between the current image and the predicted image;
- creating a transformed residual image and a transformed predicted image by transforming the residual image and the predicted image, respectively;
- obtaining a minimum threshold transform coefficient which allows a corresponding pixel of the transformed predicted image to be visually noticeable, and creating a minimum threshold image comprising the minimum threshold transform coefficient;
- and
- selecting a coefficient of the transformed residual image to be encoded from coefficients of the transformed residual image by comparing the minimum threshold transform coefficient with a corresponding coefficient of the transformed residual image.
An image encoding method comprising:
- creating a predicted image corresponding to a current image to be encoded using a prediction unit;
- creating a residual image by estimating a difference between the current image and the predicted image using a first subtraction unit;
- using a minimum threshold block creating unit for obtaining a minimum threshold coefficient which allows a corresponding pixel included in the predicted image to be visually noticeable, and creating a minimum threshold image comprising the minimum threshold coefficient;
- and
- using a second subtraction unit for creating a difference residual image by estimating a difference between the minimum threshold coefficient and a corresponding coefficient of the residual image.
An image encoding apparatus comprising:
- a prediction unit which creates a predicted image corresponding to a current image to be encoded;
- a subtraction unit which creates a residual image by estimating a difference between the current image and the predicted image;
- a first transform unit which creates a transformed residual image by transforming the residual image;
- a second transform unit which creates a transformed predicted image by transforming the predicted image;
- a minimum threshold image creating unit which obtains a minimum threshold transform coefficient which allows a coefficient included in the transformed predicted image to be visually noticeable, and creates a minimum threshold image comprising the minimum threshold transform coefficient;
- and
- a quantization unit which determines an optimum quantization step which allows a maximum quantization error to be equal to or less than the minimum threshold transform coefficient with respect to a corresponding coefficient of the transformed residual image, and quantizing the corresponding coefficient of the transformed residual image based on the determined optimum quantization step.
An image encoding apparatus comprising:
- a prediction unit which creates a predicted image corresponding to a current image to be encoded;
- a subtraction unit which creates a residual image by estimating a difference between the current image and the predicted image;
- a first transform unit which creates a transformed residual image by transforming the residual image;
- a second transform unit which creates a transformed predicted image by transforming the predicted image;
- a minimum threshold image creating unit which obtains a minimum threshold transform coefficient which allows a coefficient included in the transformed predicted image to be visually noticeable, and creates a minimum threshold image comprising the minimum threshold transform coefficient;
- and
- a selecting unit selecting a coefficient of a transformed residual image to be encoded from coefficients of the transformed residual image by comparing the minimum threshold transform coefficient with a corresponding coefficient of the transformed residual image.
An image encoding apparatus comprising:
- a prediction unit which creates a predicted image corresponding to a current image to be encoded;
- a first subtraction unit which creates a residual image by estimating a difference between the current image and the predicted image;
- a minimum threshold image creating unit which obtains a minimum threshold coefficient which allows a corresponding pixel included in the predicted image to be visually noticeable, and creates a minimum threshold image comprising the minimum threshold coefficient;
- and
- a second subtraction unit which creates a difference residual image by estimating a difference between the minimum threshold coefficient and a corresponding coefficient of the residual image.
An image decoding method comprising:
- using an entropy decoder for reading information on an optimum quantization step, which is determined such that a maximum quantization error is equal to or less than a minimum threshold transform coefficient with respect to each coefficient of a transformed residual image, and information on a coefficient of the transformed residual image, from a received bit-stream, the coefficient being quantized according to the optimum quantization step;
- inverse-quantizing the quantized coefficient of the transformed residual image using an inverse quantization unit;
- and
- restoring an image by adding a predicted image for a current image to be decoded and the inverse-transformed transformed residual image using an add unit.
An image decoding apparatus comprising:
- an entropy decoder which reads information on an optimum quantization step, which is determined such that a maximum quantization error is equal to or less than a minimum threshold transform coefficient with respect to each coefficient of a transformed residual image, and information on a coefficient of the transformed residual image, from a received bit-stream, the coefficient being quantized according to the optimum quantization step;
- an inverse-quantization unit inverse-quantizing the coefficients of the transformed residual image, which are quantized according to the read optimum quantization step;
- an inverse-transform unit which outputs a residual image by inverse-transforming the quantized coefficient of the transformed residual image;
- a prediction unit which creates a predicted image for a current image to be decoded;
- and
- an add unit which restores an image by adding the inverse-transformed residual image and the predicted image.
Dependencies are inferred from the claim text.