Block-Size Adaptive Prediction for Image Encoding and Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently processing increasing amounts of digital video data, particularly in the prediction function for current blocks, leading to suboptimal encoding and decoding efficiency and speed.
Innovation Solution
An image encoder and decoder utilize a prediction mode that includes a first prediction mode using a motion vector and a reference block in the same picture, with specific size conditions for block selection, and incorporate components like a splitter, transformer, quantizer, entropy encoder, and loop filter to enhance encoding and decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional prediction modes are used for all block sizes, then the prediction function can handle various blocks, but encoding efficiency deteriorates for small blocks
Solution Approach 1:
The patent segments the prediction mode selection based on block size. For small blocks (width ≤ 8 and height ≤ 8), a first prediction mode is applied that uses motion vectors and reference blocks from the same picture. For larger blocks, a second prediction mode is applied. This segmentation allows optimization of encoding efficiency for small blocks while maintaining versatility for all block sizes.
Solution Approach 2:
The patent implements dynamic prediction mode selection based on block dimensions. The prediction mode is not fixed but adapts dynamically according to the block size, allowing the system to switch between different prediction strategies (first prediction mode for small blocks, second prediction mode for larger blocks) to optimize encoding efficiency.
2Productivity
If a single prediction mode is used for all block sizes, then the system is simple, but encoding speed and efficiency are suboptimal
Solution Approach 1:
The patent divides the prediction process into segments based on block size thresholds. Small blocks (width ≤ 8 and height ≤ 8) use a specialized first prediction mode, while larger blocks use a second prediction mode. This segmentation enables optimized processing for different block types, improving encoding speed without excessive complexity.
Solution Approach 2:
The patent changes the prediction mode parameter based on block dimensions. By monitoring block width and height parameters, the system dynamically selects between different prediction modes, optimizing encoding speed for each block size category while maintaining manageable system complexity through clear parameter-based decision logic.
3Measurement precision
If motion vectors and reference blocks from the same picture are used, then prediction accuracy improves for small blocks, but the prediction process becomes more complex
Solution Approach 1:
The patent segments the prediction process by block size, applying the complex same-picture reference approach only to small blocks where it provides the most benefit. For larger blocks, a simpler prediction approach is used. This segmentation achieves high prediction accuracy for small blocks while limiting the complexity to only where necessary.
Solution Approach 2:
The patent applies different prediction process qualities to different block types. Small blocks receive the more complex and accurate same-picture reference prediction, while larger blocks use simpler prediction methods. This local quality differentiation optimizes prediction accuracy where it matters most without unnecessarily increasing overall system complexity.
Data Source
AI summary
An image encoder is provided including circuitry and a memory coupled to the circuitry. The circuitry, in operation, responds to a size of a block satisfying a size condition by generating a prediction image using a prediction mode selected from a plurality of prediction modes. The plurality of prediction modes include a first prediction mode in which a prediction process uses a motion vector and a reference block in a same picture as the block. The circuitry encodes the block using the prediction image.