Secondary Transform Coefficient Clipping for Image Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images/videos, particularly in immersive media formats like VR and AR, necessitates a more efficient image/video compression technique to reduce transmission and storage costs.
Innovation Solution
An image coding method and apparatus that enhances coding efficiency by employing a secondary transform process, optimizing the transformation kernel matrix, and clipping transform coefficients based on intra prediction modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution and high-quality images/videos (4K, 8K UHD) are transmitted or stored, then image quality and resolution are improved, but transmission cost and storage cost increase due to increased data amount
Solution Approach 1:
The patent extracts and removes redundant or less important information from the image data through transform coding. By converting spatial domain data to frequency domain coefficients and applying transforms like DCT or DST, the patent separates significant visual information from less important details, enabling efficient compression while maintaining perceived image quality.
Solution Approach 2:
The patent changes the representation parameters of image data by applying different transform types (DCT, DST, NSST) and quantization parameters. By dynamically selecting transform types based on prediction modes and applying adaptive quantization, the patent optimizes the balance between compression ratio and reconstruction quality, reducing data amount while preserving essential visual characteristics.
2Quantity of substance
If conventional transform coding is applied to compress image data, then data transmission and storage costs are reduced, but coding efficiency and compression performance are limited
Solution Approach 1:
The patent introduces dynamic adaptability in transform coding by dynamically selecting transform types (DCT, DST, NSST) based on intra-prediction modes and block characteristics. The transform selection is not fixed but adapts to the local image content and prediction mode, enabling optimal compression performance for different types of image regions and improving overall coding efficiency.
Solution Approach 2:
The patent employs multiple transform types with different characteristics (DCT for smooth regions, DST for edge regions, NSST for specific prediction modes) and dynamically changes the transform parameter based on the prediction mode and block type. This parameter adaptation allows the coding system to achieve better compression ratios and coding efficiency compared to conventional fixed transform approaches.
3Adaptability or versatility
If immersive media (VR, AR) and game images with different image features are compressed, then support for diverse content types is improved, but compression efficiency decreases due to varying image characteristics
Solution Approach 1:
The patent implements dynamic transform selection that adapts to different content types including VR, AR, and game images. By linking transform type selection to intra-prediction modes and analyzing local image features, the system automatically adjusts the transform strategy to match the characteristics of immersive media and game content, maintaining high compression efficiency across diverse content types.
Solution Approach 2:
The patent changes transform parameters adaptively based on content type and prediction mode. Different transform types (DCT, DST, NSST) are selected based on the specific characteristics of the content being compressed, allowing the system to optimize compression efficiency for various image features including those found in immersive media and game images with unique visual patterns.
4Measurement precision
If transform coefficients are processed without clipping, then transform precision is maintained, but computational complexity and processing overhead increase
Solution Approach 1:
The patent applies clipping operation to transform coefficients selectively, clipping only those coefficients that fall outside a predetermined range. This partial action approach maintains precision for important coefficients within the normal range while simplifying processing for extreme values, reducing computational complexity without significantly compromising reconstruction quality.
Solution Approach 2:
The patent introduces a predetermined clipping range parameter that defines the acceptable interval for transform coefficients. By applying clipping based on this parameter, the patent constrains coefficient values to a manageable range, reducing the dynamic range and computational complexity of subsequent processing steps while maintaining adequate transform precision for image reconstruction.
Data Source
AI summary
An image decoding method according to the present document comprises the steps of: receiving a quantized transform coefficient for a target block and a transform index for a non-separable secondary transform; dequantizing the quantized transform coefficients to derive transform coefficients; deriving modified transform coefficients on the basis of a matrix operation of a transform kernel matrix in a transform set indicated by the transform index and a transform coefficient list corresponding to the magnitude of dequantized transform coefficients; clipping the modified transform coefficients to a predetermined range of values; and deriving residual samples for the target block on the basis of an inverse primary transform with respect to the modified transform coefficients clipped.