Mode Dependent Intra Smoothing Lookup Table for Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression techniques face challenges in achieving high image quality without significantly increasing bit-rates, particularly due to the computational complexity of intra smoothing schemes used in video encoding and decoding processes.
Innovation Solution
A simplified Mode Dependent Intra Smoothing (MDIS) lookup table is employed, which determines whether to apply an intra smoothing filter based on prediction unit block sizes and modes, specifically turning off filtering for 8×8 pixel blocks in directional and planar prediction modes, reducing processing complexity without compromising prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intra smoothing filter is applied to all prediction units, then prediction accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent applies different smoothing filter strategies to different prediction units based on their specific characteristics. The lookup table stores filter application decisions for specific PU sizes and prediction modes, allowing the system to apply filtering only where beneficial (e.g., larger PUs or specific modes) while skipping it for others (e.g., 8×8 PUs with directional modes), thus achieving local optimization of both accuracy and complexity
Solution Approach 2:
The patent changes the parameters of the smoothing filter application by using a lookup table that stores pre-determined filter decisions based on PU size and prediction mode parameters. This allows the system to dynamically adjust whether filtering is applied based on these parameters, optimizing the balance between prediction accuracy and processing complexity for different scenarios
2Measurement precision
If mode dependent intra smoothing is implemented, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing the optimal filter application decisions in a lookup table during system initialization or standard development. This pre-computed information is then directly queried during video encoding/decoding without requiring complex real-time calculations, thus improving prediction accuracy while minimizing computational complexity during actual operation
Solution Approach 2:
The patent uses a lookup table that copies pre-determined filter decisions for different PU sizes and prediction modes. Instead of recalculating optimal filtering parameters for each prediction unit, the system copies the pre-stored decision from the LUT based on matching PU characteristics, significantly reducing computational complexity while maintaining prediction accuracy
3Measurement precision
If filtering is applied to 8×8 pixel blocks, then prediction accuracy is improved, but processing time increases
Solution Approach 1:
The patent extracts or removes the smoothing filter application for specific cases, particularly 8×8 prediction units with directional prediction modes. By taking out the filtering operation for these specific PUs while maintaining it for others, the system reduces processing time for small blocks where filtering provides minimal benefit, while still achieving good prediction accuracy overall
Data Source
AI summary
An apparatus comprising a processor configured to determine whether to apply an intra smoothing filter for a prediction unit (PU) based on a lookup table (LUT), wherein the LUT comprises data indicating the intra smoothing filter should not be applied for any PU with a block size of 8×8 pixels and associated with directional prediction mode. The disclosure also includes a method comprising generating reference samples, determining a size of a PU block, and selecting the reference samples based on PU block size, wherein filtered reference samples are not selected for PU blocks with a size of 8×8 pixels and associated with directional prediction mode.