Slice-Based Image Coding with Adaptive Probability Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image coding methods, such as the H.264 standard, face challenges in improving coding efficiency due to the need to code macroblocks in raster order while updating probability tables, which limits parallel processing and operation frequency, and fails to utilize spatial correlation between slices effectively.
Innovation Solution
An image coding method that updates probability information based on the characteristics of each region or slice, allowing for sequential coding of blocks using region-specific probability information, which is updated after coding certain blocks and before others, enabling more accurate and efficient coding by referencing spatially close blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If macroblocks are coded in raster order with continuous probability table updating, then coding efficiency is improved through adaptive probability, but parallel processing capability deteriorates and operation frequency is limited
Solution Approach 1:
The image is divided into multiple slices, and each slice is independently coded with its own probability table initialization. This segmentation allows parallel processing of different slices while maintaining adaptive probability coding within each slice, thus resolving the contradiction between coding efficiency and parallel processing capability.
Solution Approach 2:
Different probability tables are maintained for different slices, allowing each region to have optimized probability characteristics suited to its local content. This local quality approach enables parallel processing while preserving adaptive probability benefits within each slice.
2Productivity
If slices are divided for parallel coding, then parallel processing capability is improved, but spatial correlation between slices cannot be utilized and coding efficiency deteriorates
Solution Approach 1:
The probability table parameters are changed and re-initialized for each slice based on local characteristics. This allows each slice to adapt its probability parameters to local content, maintaining high coding efficiency while enabling parallel processing through slice independence.
3Manufacturing precision
If probability table is continuously updated during raster coding, then adaptive probability coding is achieved, but operation frequency improvement is limited
Solution Approach 1:
By segmenting the image into slices and re-initializing probability tables at slice boundaries, the continuous updating process is broken into discrete segments. This allows for more frequent probability table updates without continuous processing dependencies, thereby improving operation frequency while maintaining adaptive probability coding within each slice.
Data Source
AI summary
An image coding method for improving coding efficiency by using more appropriate probability information is provided. The image coding method includes: a first coding step of coding a first set of blocks included in a first region sequentially based on first probability information; and a second coding step of coding a second set of blocks included in a second region sequentially based on second probability information. In the first coding step, the first probability information is updated depending on data of a target block to be coded, after coding the target block and before coding a next target block. In the second coding step, the second probability information is updated depending on the first probability information updated in the first coding step, before coding the first target block.