Method, apparatus, and system for encoding and decoding image using LM chroma prediction
Disclosed are methods, apparatuses, and systems for encoding and decoding an image. The present invention provides an intra prediction unit that receives an input image, calculates a correlation between a Luma area block and Chroma area block in the input image in intra prediction, removes high frequency ingredients by low-pass filtering an encoded luma pixel based on the calculated correlation, and generates a prediction block by applying an LM chroma mode, which is an extended chroma mode technique, to the luma pixel removed therefrom the high frequency ingredients.
1. An image decoding apparatus, comprising:
an entropy-decoder configured to decode information specifying a Linear Model (LM) chroma prediction mode of a current block; and
an intra predictor configured to:
obtain a predicted luma block of the current block;
determine the LM chroma prediction mode among a plurality of LM chroma prediction modes specified by the decoded information;
wherein:
the plurality of LM chroma prediction modes comprises a first LM chroma prediction mode and a second LM chroma prediction mode, and
both the first LM chroma prediction mode and the second LM chroma prediction mode represent that a chroma prediction pixel in a chroma block of the current block is derived by a luma value;
obtain the luma value using a luma pixel corresponding to the chroma prediction pixel, the luma pixel being in a reconstructed luma block of the current block, and the reconstructed luma block being obtained based on the predicted luma block; and
generate the chroma prediction pixel using the luma value multiplied by a first weight,
wherein the decoded information is signaled at a block level,
wherein, under the first LM chroma prediction mode, the first weight is obtained by using a filtered luma pixel, the filtered luma pixel being obtained by using luma pixels located to a left of the reconstructed luma block,
wherein, under the second LM chroma prediction mode, the first weight is obtained without using the filtered luma pixel,
wherein the luma value is obtained by further using four neighboring luma pixels adjacent to the luma pixel corresponding to the chroma prediction pixel,
wherein the luma value is obtained by a weighted sum of the luma pixel and the four neighboring luma pixels, and
wherein a ratio between a first coefficient multiplied with the luma pixel and a second coefficient multiplied with the four neighboring luma pixels is 4:1.
2. The apparatus for decoding the image of claim 1 ,
wherein both of the first LM chroma prediction mode and the second LM chroma prediction mode are selectable for the current block regardless of color format of a current picture including the current block.
3. The apparatus for decoding the image of claim 1 ,
wherein the chroma prediction pixel is calculated by a formula as follows:
C pred [ x,y ]=α* L recon +β
([x,y] is a location of the chroma prediction pixel, α is the first weight, β is a second weight, L recon is the luma value, and C pred is the chroma prediction pixel).
4. An image encoding apparatus, comprising:
an intra predictor configured to
obtain a predicted luma block of a current block; and
an entropy-encoder configured to encode information at a block level,
wherein the information specifies a Linear Model (LM) chroma prediction mode among a plurality of LM chroma prediction modes of the current block,
wherein the plurality of LM chroma prediction modes comprises a first LM chroma prediction mode and a second LM chroma prediction mode,
wherein both the LM chroma prediction mode and the second LM chroma prediction mode represent that a chroma prediction pixel in a chroma block of the current block is derived by a luma value,
wherein the luma value is obtained by using a luma pixel corresponding to the chroma prediction pixel, the luma pixel being in a reconstructed luma block of the current block, and the reconstructed luma block being obtained based on the predicted luma block,
wherein the chroma prediction pixel is generated by using the luma value multiplied by a first weight,
wherein, under the first LM chroma prediction mode, the first weight is obtained by using a filtered luma pixel, the filtered luma pixel being obtained by using luma pixels located to a left of the reconstructed luma block,
wherein, under the second LM chroma prediction mode, the first weight is obtained without using the filtered luma pixel,
wherein the luma value is obtained by further using four neighboring luma pixels adjacent to the luma pixel corresponding to the chroma prediction pixel,
wherein the luma value is obtained by a weighted sum of the luma pixel and the four neighboring luma pixels, and
wherein a ratio between a first coefficient multiplied with the luma pixel and a second coefficient multiplied with the four neighboring luma pixels is 4:1.
5. The apparatus for encoding the image of claim 4 ,
wherein both of the first LM chroma prediction mode and the second LM chroma prediction mode are selectable for the current block regardless of color format of a current picture including the current block.
6. The apparatus for encoding the image of claim 4 ,
wherein the chroma prediction pixel is calculated by a formula as follows:
C pred [ x,y ]=α* L recon +β
([x,y] is a location of the chroma prediction pixel, α is the first weight, β is a second weight, L recon is the luma value, and C pred is the chroma prediction pixel).
7. A non-transitory computer-readable medium storing instructions that, when executed by a processor, configure the processor to:
decode information specifying a Linear Model (LM) chroma prediction mode of a current block;
obtain a predicted luma block of the current block;
obtain a reconstructed luma block based on the predicted luma block,
determine the LM chroma prediction mode among a plurality of LM chroma prediction modes specified by the decoded information, wherein:
the plurality of LM chroma prediction modes comprises a first LM chroma prediction mode and a second LM chroma prediction mode, and
both the first LM chroma prediction mode and the second LM chroma prediction mode represent that a chroma prediction pixel in a chroma block of the current block is derived by a luma value;
obtain the luma value using a luma pixel corresponding to the chroma prediction pixel, the luma pixel being in the reconstructed luma block of the current block; and
generate the chroma prediction pixel using the luma value multiplied by a first weight,
wherein the decoded information is signaled at a block level,
wherein, under the first LM chroma prediction mode, the first weight is obtained by using a filtered luma pixel, the filtered luma pixel being obtained by luma samples located to a left of the reconstructed luma block,
wherein, under the second LM chroma prediction mode, the first weight is obtained without using the filtered luma pixel,
wherein the luma value is obtained by further using four neighboring luma pixels adjacent to the luma pixel corresponding to the chroma prediction pixel,
wherein the luma value is obtained by a weighted sum of the luma pixel and four neighboring luma pixels, and
wherein a ratio between a first coefficient multiplied with the luma pixel and a second coefficient multiplied with the four neighboring luma pixels is 4:1.