Adaptive Blending Strength for Geometric Partition Mode
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding techniques face challenges in achieving high compression ratios with minimal loss in picture quality, especially when dealing with limited network resources and the need to store videos on devices with limited memory, as they struggle to efficiently manage the blending of prediction values across block partitions.
Innovation Solution
The method involves dividing a video block into sub-blocks, determining prediction modes for each, and applying a blending operation based on an indicator to combine prediction values, which is encoded in the bitstream, allowing adaptive determination of whether to perform a blending operation to avoid excessive blurring and improve subjective quality and compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a blending operation is applied to combine prediction values from multiple sub-blocks, then compression efficiency is improved and artefacts are reduced, but excessive blurring occurs and sharp edges are lost
Solution Approach 1:
The patent applies different blending strengths to different spatial locations within the block. The blending operation uses a gradient-based approach where the blending strength varies locally according to the gradient magnitude at each position. Areas with high gradient magnitudes (sharp edges) receive minimal blending to preserve edge sharpness, while areas with low gradient magnitudes (smooth regions) receive stronger blending to reduce artefacts and improve compression efficiency.
2Quantity of substance
If the video data is compressed to reduce data size for transmission and storage, then bandwidth and memory requirements are reduced, but picture quality deteriorates
Solution Approach 1:
The patent introduces a gradient-based blending strength parameter that dynamically adjusts the blending operation based on local image characteristics. By changing the blending parameter according to gradient magnitude, the system achieves better picture quality at the same compression ratio compared to uniform blending, or achieves the same picture quality at a lower compression ratio.
3Object-generated harmful factors
If a blending operation is performed for all samples in the current block, then artefacts along separation lines are eliminated, but excessive blurring occurs and subjective quality decreases
Solution Approach 1:
The patent applies different blending strengths to different spatial locations within the block. The blending operation uses a gradient-based approach where the blending strength varies locally according to the gradient magnitude at each position. Areas with high gradient magnitudes (sharp edges) receive minimal blending to preserve edge sharpness, while areas with low gradient magnitudes (smooth regions) receive stronger blending to reduce artefacts and improve compression efficiency.
Solution Approach 2:
The patent uses the gradient information copied from the original video data to control the blending operation. By copying and utilizing the gradient magnitude information, the system can make informed decisions about where to apply blending and where to preserve original details, thereby eliminating artefacts while maintaining subjective quality.
4Manufacturing precision
If adaptive blending is implemented to avoid excessive blurring, then subjective quality and compression efficiency are improved, but computational complexity increases
Solution Approach 1:
The patent applies blending selectively rather than uniformly across all blocks and samples. By performing the computationally intensive gradient calculation and adaptive blending only where necessary (in regions where it provides benefit), the system achieves improved subjective quality and compression efficiency while avoiding the full computational cost of universal adaptive blending.
Data Source
AI summary
A method for coding implemented by a decoding device or an encoding device includes: obtaining a first prediction mode for a first sub-block, obtaining a second prediction mode for a second sub-block; obtaining a first prediction value for a sample in the current block according to the first prediction mode, obtaining a second prediction value for the sample in the current block according to the second prediction mode; obtaining a value of an indicator, the value of the indicator indicating a blending operation which is performed for the sample in the current block; and obtaining a combined prediction value for the sample in the current block, according to the value of the indictor, the first prediction value, and the second prediction value. The blending operation adaptively performs around a separation line dividing the current block into at least the first sub-block and the second sub-block.