Adaptive Intra Prediction Block Sizing for 4K Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In HEVC, the maximum block size for intra prediction is limited to 32x32 pixels, which hampers coding efficiency in high-definition pictures like 4K and 8K, and 360° pictures that require more efficient intra prediction methods.
Innovation Solution
The solution involves a picture coding device and decoding device that determine block sizes based on input pictures, selecting intra prediction modes suitable for larger block sizes, and generating prediction mode candidate lists to improve prediction efficiency by using non-angular and angular intra prediction modes effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the maximum block size for intra prediction is limited to 32x32 pixels, then the prediction accuracy is maintained for standard HD pictures, but the coding efficiency deteriorates for high-definition pictures like 4K and 8K
Solution Approach 1:
The patent applies dynamics by making the block size adaptive rather than fixed. The encoder dynamically determines the block size based on the picture width and height, allowing larger blocks for high-definition pictures (4K, 8K) and 360° pictures to improve coding efficiency, while maintaining compatibility with standard HD pictures. This dynamic adjustment resolves the contradiction between coding efficiency and device complexity by allowing the system to adapt to different picture resolutions.
Solution Approach 2:
The patent changes the parameter of block size from a fixed maximum of 32x32 pixels to a variable parameter that can be determined based on picture dimensions. By allowing the block size parameter to change according to the specific picture characteristics (resolution, aspect ratio), the system achieves better coding efficiency for high-definition content without compromising the ability to handle standard HD pictures.
2Productivity
If larger block sizes are used for intra prediction, then the coding efficiency improves for high-definition pictures, but the prediction accuracy may deteriorate due to increased complexity
Solution Approach 1:
The system dynamically adjusts block sizes based on picture characteristics to optimize the balance between coding efficiency and prediction accuracy. For high-definition pictures with larger dimensions, larger blocks are used to improve coding efficiency. The dynamic determination of block sizes ensures that the system can adapt to different picture resolutions and maintain appropriate block dimensions for accurate prediction.
Solution Approach 2:
The patent applies segmentation by dividing the picture into multiple blocks with different sizes based on the picture dimensions and characteristics. This segmentation allows the system to use larger blocks in regions where they improve coding efficiency while maintaining smaller blocks where prediction accuracy is more critical, thus balancing both objectives.
3Manufacturing precision
If multiple intra prediction modes are used for larger blocks, then the prediction accuracy improves, but the processing volume increases
Solution Approach 1:
The patent changes the parameter of the number of intra prediction modes based on block size and picture characteristics. For larger blocks in high-definition pictures, the system uses a limited set of angular intra prediction modes to maintain prediction accuracy while controlling processing volume. The parameter adjustment ensures that the system does not always apply the maximum number of prediction modes, but rather selects an optimal number based on the specific picture and block characteristics.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In picture decoding device, a prediction mode candidate generation unit generates a first intra prediction mode candidate list and a second intra prediction mode candidate list from an intra prediction mode of a block neighboring a block subject to prediction. A prediction mode selection unit selects a first intra prediction mode and a second intra prediction mode from the first intra prediction mode candidate list and the second intra prediction mode candidate list, respectively. A prediction value calculation unit calculates a first prediction value from a decoded pixel neighboring the block subject to prediction, based on the first intra prediction mode, and calculates a second prediction value from a decoded pixel neighboring the block subject to prediction, based on the second intra prediction mode. A prediction value weighting unit calculates a third prediction value based on the first prediction value and the second prediction value.