BDOF Image Decoding With Time-Distance-Aware Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional BDOF processing assumes equal time distances between reference frames without considering the actual temporal distances, leading to inefficient prediction signal generation.
Innovation Solution
An image decoding device and method that sets application conditions for BDOF processing based on weight coefficients, taking into account the time distances between reference frames and the target frame, and incorporates PROF processing for gradient calculation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If BDOF processing uses equal weights for pixel values from two reference frames, then the calculation is simplified, but the prediction accuracy deteriorates because it ignores actual temporal distances between frames
Solution Approach 1:
The patent changes the weight parameters from equal weights to time-distance-dependent weights. Specifically, it introduces weight coefficients (w0, w1) that are determined based on the time distances (dt0, dt1) between the target frame and each reference frame, allowing the prediction to account for temporal variations while maintaining calculation efficiency through predefined weight tables.
Solution Approach 2:
The patent makes the weight coefficients dynamic by linking them to the time distances between frames. Instead of static equal weights, the weights adapt based on the temporal relationships, allowing the system to respond to varying frame intervals and improve prediction accuracy for different temporal scenarios.
2Productivity
If BDOF processing does not consider time distances between reference frames, then the processing is faster, but the prediction signal quality deteriorates
Solution Approach 1:
The patent performs preliminary action by pre-calculating and storing weight coefficients in a weight table based on various time distance combinations. This allows the decoding process to quickly retrieve appropriate weights without performing complex real-time calculations, thus maintaining high processing speed while improving prediction quality through time-distance-aware weighting.
Solution Approach 2:
The patent replaces the mechanical calculation process with a lookup-based system. Instead of computing weights dynamically during decoding, the system substitutes this with a table lookup operation that provides the same functional result with reduced computational overhead, maintaining productivity while enhancing prediction accuracy.
3Device complexity
If the application condition of BDOF does not include time distance conditions, then the processing is simpler, but the adaptability to different temporal scenarios deteriorates
Solution Approach 1:
The patent modifies the application condition parameters by introducing time distance thresholds (th0, th1) that determine whether BDOF processing should be applied. This allows the system to adapt to different temporal scenarios by comparing actual time distances against these thresholds, enhancing versatility while maintaining relatively simple decision logic through predefined threshold values.
Data Source
AI summary
An image decoding device includes: a prediction signal generation unit configured to determine whether or not an application condition of Bi-Directional Optical Flow (BDOF) processing is satisfied for each block, generate a prediction signal by executing the BDOF processing in a case where it is determined that the application condition is satisfied, and set the application condition based on a weight coefficient in a case where calculation is performed using pixel values of two reference frames or values calculated from the pixel values in the BDOF processing such that the application condition includes a condition on time distances between the two reference frames and a target frame.