Image Decoding Partition Prediction Using Neighboring Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image compression techniques face challenges in efficiently coding and decoding partition information for target blocks, leading to increased data requirements and reduced image compression efficiency.
Innovation Solution
The method involves obtaining and predicting partition information from neighboring blocks, coding the difference between actual and predicted partition information, and transmitting this data to reduce the number of bits needed for coding, while improving compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If partition information is coded directly without prediction, then coding simplicity is maintained, but the number of bits required increases and compression efficiency decreases
Solution Approach 1:
The patent applies preliminary action by predicting partition information before actual coding. The decoder predicts partition information of the current block using partition information of neighboring blocks before decoding the actual partition information, allowing the system to code only the difference between predicted and actual values, thereby reducing bit requirements while maintaining manageable complexity through pre-computed predictions
2Measurement precision
If partition information of all blocks is transmitted in full, then decoding accuracy is ensured, but data transmission volume increases
Solution Approach 1:
The patent applies the extraction principle by removing redundant information from the data stream. Since partition information can be predicted from neighboring blocks with high accuracy, the system extracts and transmits only the difference between predicted and actual partition information, significantly reducing data volume while maintaining decoding accuracy through the combination of predicted values and small correction differences
3Manufacturing precision
If block partitioning is performed to improve picture quality, then image quality increases, but the amount of partition information to be coded increases
Solution Approach 1:
The patent applies merging by combining the partition information coding process with neighboring block information. Instead of treating each block's partition information independently, the system merges the coding of current block partition information with the already-coded partition information of neighboring blocks, using the neighboring information as a prediction basis to reduce the complexity of coding partition information for finely partitioned blocks
Data Source
AI summary
The present invention relates to image encoding and image decoding. A method for encoding images according to the present invention comprises the steps of: acquiring division information from peripheral blocks of an encoding target block with respect to the peripheral blocks; predicting the division of the encoding target block on the basis of the acquired division information; encoding information on the difference between the practical division information of the encoding target information and the predictive division information predicted in the predicting process; and transmitting the encoded information. The present invention can encode/decode the division information of the encoding/decoding target blocks using the information about the peripheral blocks.