Intra Prediction Using Dual Reference Samples for Image Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images, such as HD and UHD, leads to higher image data volumes, resulting in increased transmission and storage costs due to the need for more bits, necessitating a highly efficient image compression technique.
Innovation Solution
The method and device enhance image coding efficiency by performing intra-prediction based on first and second neighboring samples of a current block, determining the reconstruction order of transform units (TUs) within a prediction unit (PU) based on the intra-prediction mode, and deriving non-square TUs from a current coding unit (CU), thereby improving prediction accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution and high-quality images are transmitted or stored, then image quality is improved, but transmission cost and storage cost increase due to increased data volume
Solution Approach 1:
The patent extracts and utilizes only the necessary reference samples (first and second neighboring samples) required for intra-prediction, discarding redundant information. This selective extraction maintains image quality while reducing the amount of data that needs to be processed and transmitted.
Solution Approach 2:
The patent segments the prediction process into distinct stages: deriving first neighboring samples, deriving second neighboring samples from the first, and then using these segmented sample sets for prediction. This segmentation allows efficient processing of reference data while maintaining prediction accuracy.
2Device complexity
If conventional intra-prediction using only first neighboring samples is used, then device complexity is reduced, but prediction accuracy deteriorates
Solution Approach 1:
The patent performs preliminary action by deriving first neighboring samples before deriving second neighboring samples. This sequential preparation ensures that all necessary reference data is ready before the actual prediction process, improving accuracy without requiring complex simultaneous processing.
Solution Approach 2:
The patent extends the prediction approach by adding a second dimension of reference samples (second neighboring samples derived from first neighboring samples) to the traditional single-dimension approach. This dimensional expansion provides more comprehensive reference data for improved prediction accuracy.
3Speed
If fixed reconstruction order of TUs is used, then processing speed is maintained, but prediction accuracy for different intra-prediction modes deteriorates
Solution Approach 1:
The patent introduces dynamic adaptability by allowing the reconstruction order of TUs to change based on the intra-prediction mode. Different scanning orders (e.g., raster scan, zigzag scan) are selected dynamically according to the prediction requirements, optimizing both speed and accuracy for different scenarios.
4Device complexity
If square TUs are used for all CUs, then device complexity is reduced, but prediction accuracy for non-square regions deteriorates
Solution Approach 1:
The patent applies asymmetry by allowing TUs to have different shapes (square or non-square) based on the characteristics of the current CU and the intra-prediction mode. This asymmetric approach matches the geometric properties of the image content, improving prediction accuracy for non-square regions without requiring complex processing for all cases.
Data Source
AI summary
An image decoding method performed by a decoding device according to the present invention comprises the steps of: deriving an intra prediction mode for a current PU on the basis of a received bit stream; and performing reconstruction of a plurality of TUs located in an area of the current PU on the basis of the intra prediction mode, wherein the step of performing reconstruction comprises, with respect to a current block that is one of the plurality of TUs, a step of deriving a prediction sample of the current block on the basis of a first reference sample located in a prediction direction of the intra prediction mode and a second reference sample located in an opposite direction of the prediction direction, and a step of reconstructing a block to be reconstructed on the basis of the prediction sample.