Pixel-Adaptive Filters for Low-Bit-Rate Frame Restoration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies use fixed filters that do not adapt to the specific characteristics of individual frames, leading to suboptimal decoding quality and inefficiencies in data processing and storage.
Innovation Solution
Implement pixel-adaptive filters obtained by combining a filter bank using pixel-specific combining scalars, derived from local pixel information and potentially aided by side-information, to enhance frame restoration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If fixed filters are used for decoding, then device complexity is reduced, but decoding quality deteriorates due to inability to adapt to frame-specific characteristics
Solution Approach 1:
The patent applies dynamics by transitioning from fixed filters to adaptive filters that change based on frame characteristics. The filter parameters are dynamically adjusted using motion compensation techniques, where the filter for the current frame is derived from the filter of the previous frame combined with motion vector information, allowing the system to adapt to temporal variations in video content while maintaining computational efficiency
Solution Approach 2:
The patent implements local quality by applying different filters to different regions of the frame based on local motion characteristics. Instead of using a single global filter, the system computes region-specific filters using motion-compensated approaches, where each region's filter adapts to its local motion patterns, thereby improving decoding quality in heterogeneous regions with diverse motion behaviors
2Manufacturing precision
If pixel-adaptive filters are used, then decoding quality improves, but computational complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing filters for reference frames during the encoding phase. These pre-computed filters are then reused and adapted for current frames through motion compensation, avoiding the need to compute filters from scratch for every frame. This approach significantly reduces computational complexity while maintaining the benefits of adaptive filtering
Solution Approach 2:
The patent implements copying by reusing filters from reference frames and copying them to current frames with minimal modifications based on motion information. Instead of independently computing pixel-adaptive filters for each frame, the system copies and adapts existing filters using motion vector offsets, thereby reducing computational load while preserving the adaptive filtering benefits
3Productivity
If lossy encoding techniques are used, then data transmission efficiency improves, but image quality deteriorates due to encoding artifacts
Solution Approach 1:
The patent converts the harmful effect of lossy encoding artifacts into a benefit by designing filters specifically tailored to remove the types of artifacts introduced by compression. The adaptive filters learn and compensate for the characteristic artifacts produced by quantization and block-based encoding, transforming the inevitable quality loss from lossy compression into an opportunity for targeted restoration and enhancement
Data Source
AI summary
Obtaining a restored frame from a degraded frame includes obtaining, for a pixel of the degraded frame, magnitude features based on a first window centered at the pixel. A cardinality N of the magnitude features is at least 1. The magnitude features are used to obtain a pixel-adaptive filter. The pixel-adaptive filter is applied to the pixel to obtain a pixel of the restored frame.