Combined Intra Prediction for Higher Video Compression Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression technologies face challenges in efficiently leveraging intra prediction modes to reduce redundancy in image and video coding, leading to suboptimal compression efficiency.
Innovation Solution
A method and apparatus for encoding and decoding video blocks by combining two distinct intra prediction modes, such as Planar and directional modes, using weighted averages based on reconstructed samples and decoder-side derivations, to enhance compression performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single intra prediction mode is used for encoding video blocks, then the encoding process is simple and fast, but the compression efficiency is suboptimal due to inability to leverage multiple prediction approaches
Solution Approach 1:
The patent combines multiple intra prediction modes (planar mode and directional modes) into a unified prediction framework. The encoder generates predictions from both planar and directional modes, then merges them using weighted averaging to create a final prediction signal. This merging approach allows the system to leverage the strengths of different prediction methods simultaneously, improving compression efficiency by better capturing spatial redundancy in video content.
2Loss of information
If multiple intra prediction modes are combined to improve compression efficiency, then redundancy reduction is enhanced, but the encoding and decoding complexity increases
Solution Approach 1:
The patent implements a selective combination strategy where not all prediction modes are combined in every case. The encoder evaluates whether combining planar and directional modes provides sufficient benefit to justify the additional complexity. By applying combination only when beneficial (partial action) rather than universally (excessive action), the system achieves improved redundancy reduction while controlling encoding complexity through conditional application of the combined mode approach.
Data Source
AI summary
A method and an apparatus for encoding or decoding a block of an image or a video are disclosed. For at least one block of an image or a video, a first intra prediction mode is obtained, a second intra prediction mode is obtained. The at least one block is encoded or decoded based on a combination of the first intra prediction mode and the second intra prediction mode. In an embodiment, the first intra prediction mode is a non-direction based intra prediction mode, the second intra prediction mode is a directional intra prediction mode.