CIIP Image Decoding With Planar Intra Prediction for Lower Bitrate
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high resolution and quality images leads to higher data volumes, increasing transmission and storage costs, necessitating improved image encoding/decoding techniques.
Innovation Solution
Implementing combined inter intra prediction (CIIP) for both luma and chroma blocks, using intra and inter prediction values weighted in a Planar mode, with specific conditions for CIIP mode information based on block size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high resolution and quality images are used, then image quality is improved, but data volume increases leading to higher transmission and storage costs
Solution Approach 1:
The patent combines inter prediction and intra prediction into a unified CIIP framework, merging two separate prediction mechanisms to work together. This allows the encoder to leverage both temporal (inter) and spatial (intra) redundancies simultaneously, achieving superior compression efficiency for high resolution images while maintaining image quality.
Solution Approach 2:
The patent dynamically adjusts the weighting parameters (w0, w1) for combining intra and inter prediction values based on block characteristics, motion complexity, and prediction mode. This parameter optimization enables adaptive compression that maintains high image quality while minimizing data volume for different regions of the image.
2Productivity
If combined inter intra prediction (CIIP) is applied to both luma and chroma blocks, then encoding efficiency is improved, but device complexity increases
Solution Approach 1:
The patent applies CIIP selectively to different block types (luma and chroma) and different regions within blocks based on local characteristics such as motion complexity, texture patterns, and prediction mode. This localized application optimizes encoding efficiency for each region while avoiding unnecessary complexity in regions where simple prediction suffices.
Solution Approach 2:
The patent divides the image into multiple blocks and applies different prediction strategies (CIIP, intra-only, inter-only) to different blocks based on their characteristics. This segmentation allows the encoder to manage complexity by processing each block independently with the most appropriate method, rather than applying a uniform complex approach to the entire image.
3Measurement precision
If CIIP mode information is acquired from bitstream for all block sizes, then decoding accuracy is improved, but transmission overhead increases
Solution Approach 1:
The patent acquires CIIP mode information from the bitstream selectively based on block size conditions. For blocks meeting specific size criteria (e.g., larger blocks where CIIP provides significant benefit), the mode information is explicitly transmitted. For smaller blocks or blocks where CIIP provides minimal improvement, the decoder infers the mode or uses default behavior, reducing transmission overhead while maintaining adequate decoding accuracy.
Data Source
AI summary
Disclosed herein is an image encoding/decoding method and apparatus. The image decoding method includes deriving an intra prediction value by performing intra prediction with respect to a current block, deriving an inter prediction value by performing inter prediction with respect to the current block, deriving a combined inter intra prediction (CIIP) value by weighting the intra prediction value and the inter prediction value, and generating a reconstructed block of the current block using the CIIP value. The intra prediction is performed in a Planar mode.