Mixed Domain In-Loop Filter for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding methods, such as those using hybrid codecs like H.264/AVC and H.265/HEVC, face challenges in improving prediction signals and decoded video quality due to quantization noise and sensor noise, with current in-loop filters like the Adaptive Loop Filter having limitations in efficiency and complexity.
Innovation Solution
A video coding apparatus and method that jointly determines and applies filter parameters in both frequency and pixel domains within a loop, using parameters based on the reconstructed frame and codec signaling information, allowing for improved filtering without requiring the original frame and reducing the need for parameter transmission between encoder and decoder.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If Adaptive Loop Filter is used with parameter estimation block, then filtering precision is improved, but device complexity increases due to multiple blocks and parameter transmission
Solution Approach 1:
The patent combines the parameter estimation block and pixel domain filtering block into a single integrated filter. This merging eliminates the need for separate parameter estimation and transmission between encoder and decoder, reducing device complexity while maintaining filtering precision through the unified structure that directly processes reconstructed frames.
Solution Approach 2:
The integrated filter operates autonomously on the decoder side using only the reconstructed frame as input. It performs self-service by estimating parameters internally during the filtering process without requiring external parameter transmission, thereby simplifying the overall system architecture while preserving filtering effectiveness.
2Manufacturing precision
If filter parameters are transmitted from encoder to decoder, then filtering precision is improved, but loss of information increases due to bitstream overhead
Solution Approach 1:
The decoder-side integrated filter performs self-service by internally estimating the parameters needed for filtering directly from the reconstructed frame. This eliminates the need to transmit filter parameters through the bitstream, reducing information loss and bitstream overhead while maintaining filtering precision through autonomous parameter estimation.
3Device complexity
If single filter is used for both in-loop and post filtering, then device complexity is reduced, but filtering precision deteriorates compared to separate filters
Solution Approach 1:
The integrated filter incorporates dynamic adaptability by allowing different filtering operations to be applied to different regions of the reconstructed frame. The filter can dynamically adjust its behavior based on local characteristics, enabling it to perform both in-loop and post-filtering functions with precision comparable to separate filters while maintaining lower device complexity.
Data Source
AI summary
A video coding apparatus is described for encoding or decoding a frame of a video, the video coding apparatus comprising a computer program code and a processor configured to carry out operations on the frame of video. The operations include reconstructing the frame, and determining one or more filter parameters and filtering in a frequency domain and a pixel domain a block of the reconstructed frame based on the determined filter parameters to obtain a filtered block of the frame. The determining the one or more filter parameters and filtering in a frequency domain based on the determined filter parameters is performed jointly within a loop, and the one or more filter parameters are based on one or more first parameters that are based on the reconstructed frame and one or more second parameters which are based on a codec signaling information.