Apparatus and method for image encoding/decoding considering impulse signal

1. A method for video encoding, the method comprising:

generating a predicted block by predicting a current block;

generating an M×N residual block by subtracting the predicted block from the current block; when the M×N residual block includes no impulsive components having values larger than the predetermined value, setting all residual signals of the M×N residual block to zero; and

when one or more of frequency coefficients generated through transformation and quantization of the M×N residual block are not zero, and

when the M×N residual block contains one or more impulsive components having values larger than a predetermined value, performing sub-steps of:

selecting an A×B residual block containing the impulsive components and then transforming the A×B residual block into frequency coefficients, wherein wherein the A×B residual block is selected such that the A×B residual block is smaller in size than the entire M×N residual block; and the number of pixels in the A×B residual block is more than a predetermined number,

encoding the frequency coefficients of the A×B residual block into a bitstream, and

encoding information on a position and a size of the A×B residual block into the bitstream,

wherein residual signals in the M×N residual block, which are excluded from the A×B residual block, are not encoded into the bitstream.

2. The method of claim 1 , wherein the performing the sub-steps further comprises:

encoding, into the bitstream, identification information indicating that not the M×N residual block but the A×B residual block is encoded.

3. The method of claim 1 , wherein the encoding the frequency coefficients of the A×B residual block further comprises differently scanning frequency coefficients generated through transformation and quantization of the A×B residual block, according to the shape of the A×B residual block.

4. The method of claim 1 , wherein the A×B residual block is shaped as a rectangular block or a diagonal arrangement of blocks.

5. The method of claim 4 , wherein the performing the sub-steps further comprises:

encoding diagonal block identification information into the bitstream when the A×B residual block is shaped as the diagonal arrangement of blocks.

6. The method of claim 1 , wherein the A×B residual block is selected to have an impulse rate that is higher than a threshold value.

7. The method of claim 1 , wherein the performing the sub-steps further comprises:

encoding, into the bitstream, identification information indicating the encoding of the A×B residual block.

8. The method of claim 1 , wherein the performing the sub-steps further comprises:

inserting replacement residual signal information into the bitstream, wherein the replacement residual signal information identifies a replacement residual value for all residual signals of the M×N residual block that are not in the A×B residual block.

9. An apparatus for video encoding, the apparatus comprising:

a predictor configured to generate a predicted block which is predicted from a current block;

a subtractor configured to generate an M×N residual block by subtracting the predicted block from the current block; an encoding mode determiner configured to set all residual signals of the M×N residual block to zero, when the M×N residual block includes no impulsive components having values larger than the predetermined value; and

an A×B encoder configured to, in response to the M×N residual block containing one or more impulsive components having values larger than a predetermined value, when one or more of frequency coefficients generated through transformation and quantization of the M×N residual block are not zero,

select an A×B residual block containing the impulsive components and then transform the A×B residual block into frequency coefficients, wherein the A×B residual block is selected such that the A×B residual block is smaller in size than the entire M×N residual block and the number of pixels in the A×B residual block is more than a predetermined number;

encode the frequency coefficients of the an A×B residual block into a bitstream, and

encode information on a position and a size of the A×B residual block into the bitstream,

wherein residual signals in the M×N residual block, which are excluded from the A×B residual block, are not encoded into the bitstream.

10. The apparatus of claim 9 , further comprising:

an M×N encoder configured to encode the M×N residual block to generate the bitstream,

an wherein the encoding mode determiner is further configured to determine between the M×N encoder and the A×B encoder to encode, depending on whether the residual signals which include the one or more impulsive components are included in the M×N residual block.

11. The apparatus of claim 9 , wherein the A×B encoder is further configured to

encode, into the bitstream, identification information indicating that not the M×N residual block but the A×B residual block is encoded.

12. A method for video decoding, the method comprising: identifying, from a bitstream, whether all quantized frequency coefficients of an M×N residual block are zero; when all quantized frequency coefficients of the M×N residual block are zero, setting all residual signals of the M×N residual block to zero; when one or more of quantized frequency coefficients of the M×N residual block are not zero, performing sub-steps of:

decoding the bitstream to extract information on a position and a size of an A×B residual block which is a part of the M×N residual block wherein the number of pixels in the A×B residual block is more than a predetermined number,

decoding the bitstream to extract frequency coefficients of the A×B residual block according to the extracted information,

generating the A×B residual block by using the frequency coefficients of the A×B residual block, and;

generating the M×N residual block by

inserting the A×B residual block into a corresponding position in the M×N residual block, based on the extracted information, and

adding one or more residual signals into remaining positions in the M×N residual block but outside the A×B residual block;

generating a predicted block by predicting a current block; and

reconstructing the current block by adding the predicted block to the generated M×N residual block,

wherein the one or more residual signals are set to a predetermined value or a replacement residual value extracted from the bitstream.

13. The method of claim 12 , wherein the predetermined value is zero.

14. An apparatus for video decoding, the apparatus comprising:

a decoder configured to identify, from a bitstream, whether all quantized frequency coefficients of an M×N residual block are zero, and

when one or more of quantized frequency coefficients of the M×N residual block are not zero, decode the bitstream to extract information on a position and a size of the A×B residual block which is a part of the M×N residual block and extract frequency coefficients of the A×B residual block according to the extracted information, wherein the number of pixels in the A×B residual block is more than a predetermined number;

an A×B residual block generator configured to generate the A×B residual block by using the frequency coefficients of the A×B residual block;

an M×N residual block generator configured to when all quantized frequency coefficients of the M×N residual block are zero, setting all residual signals of the M×N residual block to zero, and when one or more of quantized frequency coefficients of the M×N residual block are not zero, generate the M×N residual block by

inserting the A×B residual block into a corresponding position in the M×N residual block based on the extracted information, and

adding one or more residual signals into remaining positions in the M×N residual block but outside the A×B residual block;

a predictor configured to generate a predicted block by predicting from a current block; and

an adder configured to reconstruct the current block by adding the predicted block to the generated M×N residual block,

wherein the one or more residual signals are set to a predetermined value or a replacement residual value extracted from the bitstream.