CNN-Based Video Decoding Filter for Quantization and Blocking Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing data volume and demand for high-capacity video content necessitate more efficient video encoding and decoding techniques, particularly in addressing quantization errors and blocking artifacts.
Innovation Solution
Applying a convolutional neural network (CNN)-based filter to video decoding operations, utilizing quantization parameter maps and block partition maps to enhance reconstructed pictures and improve prediction accuracy through supervised learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a CNN-based filter is applied to video decoding to reduce quantization errors and blocking artifacts, then image quality is improved, but decoding complexity increases
Solution Approach 1:
The CNN filter is trained offline in advance using supervised learning with ground truth images. The trained filter parameters are then stored and directly applied during decoding without requiring real-time training, thus improving image quality while keeping runtime complexity manageable
Solution Approach 2:
Instead of implementing a full CNN architecture during decoding, the patent uses a simplified version that copies the essential filtering function. The complex training process is replaced by applying pre-computed filter kernels during the actual decoding operation
2Productivity
If video data is compressed using traditional encoders, then data transmission efficiency is improved, but quantization errors and blocking artifacts increase
Solution Approach 1:
The patent converts the harmful effects of quantization (which causes blocking artifacts and image degradation) into a learning opportunity. The CNN filter is trained specifically to recognize and correct these quantization-induced artifacts, transforming the previously harmful compression effects into identifiable patterns that can be systematically corrected during decoding
Solution Approach 2:
The CNN-based filter acts as an intermediary between the compressed video data and the final reconstructed image. It processes the decoded video data to remove quantization errors and blocking artifacts, serving as a intermediate correction step that improves image quality without requiring changes to the original compression algorithm
Data Source
AI summary
The present disclosure relates to video encoding or decoding and, more specifically, to an apparatus and a method for applying an artificial neural network (ANN) to video encoding or decoding. The apparatus and the method of the present disclosure are characterized by applying a CNN-based filter to a first picture and at least one of a quantization parameter map and a block partition map to output a second picture.