Image Decoding with Size-Based Transform Block Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images leads to a significant increase in transmission and storage costs due to the large amount of information required, necessitating a more efficient image compression technique.
Innovation Solution
The method involves partitioning blocks based on size and transform kernel, and performing transforms and deblocking filtering on these partitioned blocks, minimizing partitioning information and reducing computation complexity while improving image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image data is transmitted using conventional wired/wireless broadband lines or stored using existing storage media, then high-resolution and high-quality images can be transmitted and stored, but the transmission cost and storage cost increase significantly
Solution Approach 1:
The image is divided into blocks of specific sizes (e.g., 4x4, 8x8, 16x16 pixels) that can be processed and transformed independently. This segmentation allows the image data to be compressed more efficiently by applying transform operations to smaller units, reducing the overall bit rate while maintaining quality.
Solution Approach 2:
The patent applies transform operations (such as discrete cosine transform or discrete sine transform) to convert spatial frequency data into transform coefficients. This parameter transformation enables better compression by concentrating energy in fewer coefficients, allowing for more efficient transmission and storage.
2Productivity
If the block size is increased to improve coding efficiency, then compression performance improves, but the transform kernel size constraint cannot be satisfied
Solution Approach 1:
When the block size exceeds the maximum transform kernel size, the block is divided into multiple smaller transform blocks. This segmentation ensures that each transform operation works on a size compatible with the transform kernel constraints, while still processing large blocks efficiently through multiple smaller units.
Solution Approach 2:
The patent introduces a hierarchical block structure where blocks can be partitioned into multiple transform blocks at different levels. This dimensional approach allows the system to handle various block sizes by recursively dividing them into manageable units that satisfy the transform kernel size constraints.
3Device complexity
If block partitioning is performed to satisfy transform kernel size constraints, then transform operations can be performed, but additional partitioning information must be transmitted
Solution Approach 1:
The block partitioning structure is derived automatically from the block size and transform kernel size constraints, without requiring explicit partitioning information to be transmitted. The decoder can independently determine the partitioning scheme by applying the same constraints, eliminating the need to send additional partitioning data.
Solution Approach 2:
The transform block structure serves multiple functions: it satisfies the transform kernel size constraints, provides the basis for transform operations, and implicitly defines the processing units for compression. This multi-functionality reduces the need for separate partitioning information transmission.
Data Source
AI summary
A method for decoding an image by a decoding apparatus according to the present invention comprises the steps of: deriving predicted samples by performing prediction on a current block; determining whether the current block is to be divided, on the basis of a size and a maximum transform size of the current block; when the current block is determined to be divided, dividing the current block into transform blocks; deriving transform coefficients of the transform blocks; deriving residual samples by performing an inverse transform procedure for the transform coefficients; and generating a reconstructed picture on the basis of the residual samples and the predicted samples, wherein, when the size of the current block is larger than the maximum transform size, the current block is determined to be divided.