Image Filtering Using Neighboring Block Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image coding techniques face challenges in improving image coding efficiency and subjective/objective quality, particularly in adapting image filtering methods using previously decoded information before the decoding process of a current block.
Innovation Solution
The proposed solution involves deriving filter information for a current block based on previously decoded neighboring blocks or samples, which reduces the need for transmitting filter data and allows for adaptive filtering, enhancing coding efficiency and image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If filter information is derived independently at encoder and decoder sides through template-based procedure, then filtering accuracy is improved, but data transmission requirements increase and coding efficiency decreases
Solution Approach 1:
The patent applies preliminary action by deriving filter information at the encoder side before encoding and transmitting it in the bitstream. The decoder then uses this pre-derived filter information directly without needing to perform independent template-based derivation, thereby reducing data transmission requirements while maintaining filtering accuracy.
Solution Approach 2:
The patent uses the encoded bitstream as an intermediary to transfer filter information from encoder to decoder. Instead of transmitting raw filter parameters directly, the system encodes them in a compressed format that reduces transmission overhead while preserving the necessary filtering characteristics for accurate reconstruction.
2Manufacturing precision
If adaptive filter is determined through template-based procedure, then image quality is improved, but device complexity increases
Solution Approach 1:
The patent extracts the complex template-based filter determination process from the decoder side and relocates it to the encoder side. This extraction reduces the computational complexity and processing requirements at the decoder, while the encoder performs the intensive template matching and filter selection operations during the encoding phase.
Solution Approach 2:
The complex adaptive filter determination through template-based procedure is performed in advance at the encoder side before transmission. The decoder receives pre-determined filter information that requires minimal processing, thereby reducing device complexity at the decoding end while maintaining high image quality through the sophisticated filter selection.
3Productivity
If previously decoded neighboring information is used for filtering current block, then coding efficiency is improved, but filtering adaptability may be reduced
Solution Approach 1:
The patent implements feedback by using previously decoded neighboring blocks as reference information for determining the filter parameters of the current block. The decoder leverages the already-decoded spatial context to adaptively select appropriate filters, creating a feedback loop that improves coding efficiency while maintaining filtering adaptability through context-aware decision making.
Solution Approach 2:
The patent changes the parameters used for filter determination by incorporating previously decoded neighboring block information into the filter selection process. Instead of relying solely on template matching, the system adjusts filter parameters based on the actual decoded context, thereby improving coding efficiency while preserving adaptability through dynamic parameter adjustment.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A filtering method performed by a decoding device according to the present invention comprises the steps of: deriving a specific neighboring block of a current block so as to derive filter information on the current block; deriving the specific neighboring block and a reference block within a reference picture corresponding to the specific neighboring block; deriving the filter information on the current block on the basis of the specific neighboring block and the reference block; and filtering the current block on the basis of the filter information. According to the present invention, the overall coding efficiency can be improved by deriving the filter information on the basis of the neighboring block of the already decoded current block and reducing or removing the amount of data for the transmission of the information necessary for filter generation.