Intra-Prediction Block Size Selection for Image Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image coding methods, such as those using the H.264 standard, face challenges in suppressing the amount of generated code without significantly deteriorating image quality, especially when dealing with input images featuring complex patterns like checked patterns, where simply increasing the quantization parameter fails to effectively reduce code overhead.
Innovation Solution
An image coding apparatus that performs intra-prediction coding in units of multiple block sizes, determining the optimal block size based on statistical information of pixel values, such as variance and edge detection, to select the most appropriate prediction mode for reducing code generation while maintaining image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the quantization parameter is increased to suppress the amount of generated code, then the code amount is reduced, but image quality deteriorates significantly
Solution Approach 1:
The patent changes the parameter of prediction block size from a fixed value to a variable that can be selected from multiple sizes (4×4, 8×8, 16×16 pixels). By adapting the block size to the local image characteristics (smooth regions use larger blocks, detailed regions use smaller blocks), the system achieves better prediction accuracy without relying solely on quantization, thus reducing code amount while preserving image quality
Solution Approach 2:
The patent segments the macroblock into multiple prediction blocks of different sizes based on image characteristics. Instead of applying a uniform quantization parameter across the entire block, the system divides the block and applies different prediction block sizes to different regions, allowing fine-grained control over the balance between code amount and image quality in each segment
2Quantity of substance
If the prediction block size is increased to reduce overhead information, then the code amount is reduced, but prediction accuracy fails for complex patterns like checked patterns
Solution Approach 1:
The patent introduces dynamic selection of prediction block sizes based on local image characteristics. The system calculates variance or gradient metrics for different regions and adaptively chooses the appropriate block size (4×4, 8×8, or 16×16) for each region. This dynamic adaptation allows the system to use larger blocks in smooth regions to reduce overhead while using smaller blocks in complex regions to maintain prediction accuracy
Solution Approach 2:
The patent applies different prediction block sizes to different local regions of the macroblock based on their specific characteristics. Smooth regions use larger 16×16 blocks to minimize overhead, while regions with edges or patterns use smaller 4×4 or 8×8 blocks to maintain prediction accuracy. This local optimization ensures that each region is processed with the most appropriate block size for its content
3Manufacturing precision
If multiple prediction block sizes are used to improve prediction performance, then image quality is improved, but the complexity of determining the optimal block size increases
Solution Approach 1:
The patent implements self-service through automated variance calculation and block size selection. The system automatically computes the variance of pixel values in each region and uses this metric to determine the optimal prediction block size without requiring manual intervention or complex external control mechanisms. This self-service approach manages the complexity internally while maintaining simple interfaces
Solution Approach 2:
The patent performs preliminary calculation of variance or gradient metrics for each potential prediction block region before finalizing the block size selection. By pre-computing these characteristics, the system prepares the necessary information in advance, making the subsequent block size determination more efficient and less complex during the actual encoding process
Data Source
AI summary
An image coding apparatus and an image coding method allow selecting, according to an image in a macroblock to be coded, a most appropriate intra prediction mode for controlling the amount of generated code. The image coding apparatus includes a block feature quantity calculator which calculates, based on pixel values of pixels in the current macroblock in an input image, statistical information of each of the pixel values; and an intra-prediction block size determiner which determines, based on the calculated statistical information, an intra-prediction block size according to a predetermined reference such that a larger intra-prediction block size is more likely to be selected when a degree of change in the pixel value corresponding to a predetermined direction in the current macroblock is smaller. A encoder which performs intra-prediction coding on the current macroblock, in units of the intra-prediction block having the determined size.