Weighted Prediction Estimation Using Displaced Frame Differential
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression standards, such as JVT (H.264 and MPEG AVC), face challenges in accurately determining weighted prediction parameters for efficient video encoding and decoding, particularly in handling fading sequences and bi-predictive macroblocks.
Innovation Solution
A method for weighted prediction estimation using a displaced frame differential, where a video encoder computes motion vectors, motion compensates reference pictures, applies weighting factors, and encodes the difference between the uncompressed image block and the weighted motion compensated reference picture, along with the corresponding index of the reference picture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional video compression standards (MPEG-1, 2, 4) use single reference picture prediction without scaling, then the encoding process is simple, but the prediction accuracy deteriorates for fading sequences and bi-predictive macroblocks
Solution Approach 1:
The patent applies parameter changes by introducing weighting factors and offsets that modify the prediction parameters dynamically. Instead of using fixed equal weighting (1/2, 1/2) for bi-prediction, the invention allows arbitrary multiplicative weighting factors and additive offsets to be applied to reference picture predictions, enabling accurate representation of fading sequences and varying prediction quality across different macroblocks
Solution Approach 2:
The invention implements dynamics by allowing different macroblocks in the same picture to use different weighting factors through reference picture reordering and memory management control operation (MMCO) commands. This dynamic adaptation enables the encoding system to adjust prediction parameters locally based on scene characteristics, improving accuracy for fading sequences while maintaining flexibility in the encoding process
2Measurement precision
If JVT standard uses explicit mode with coded weighting factors and offsets, then the prediction accuracy improves, but the bit rate increases due to additional parameters to be transmitted
Solution Approach 1:
The patent employs feedback mechanisms where the encoder and decoder use the same MMCO commands and reference picture reordering logic to ensure consistent interpretation of reference picture indices. This feedback alignment allows the system to achieve accurate weighted prediction while minimizing the information that needs to be transmitted, as both encoder and decoder independently derive the same weighting parameters from shared control data
Solution Approach 2:
The invention applies preliminary action by pre-establishing the reference picture list ordering and MMCO command sequences at the encoder, which then guide the decoder's reconstruction process. The weighting factors and offsets are determined based on pre-negotiated reference picture arrangements, allowing accurate prediction without transmitting all parameter details explicitly
3Quantity of substance
If implicit mode is used for bi-prediction in B slices, then the bit rate is reduced by not transmitting weighting parameters, but the adaptability to different scene conditions deteriorates
Solution Approach 1:
The patent achieves universality by designing the MMCO command structure to serve multiple functions: it manages reference picture list ordering, controls memory management, and implicitly defines weighting parameters for both explicit and implicit modes. This multi-functional approach allows the same control mechanism to adapt to different scene conditions (fading sequences, static scenes, motion variations) while maintaining a compact bit rate representation
Data Source
AI summary
A video encoder and corresponding methods are provided for weighted prediction estimation using a displaced frame difference or differential, the encoder embodying a method for receiving a substantially uncompressed image block, assigning a weighting factor for the image block corresponding to a particular reference picture having a corresponding index, computing motion vectors corresponding to the difference between the image block and the particular reference picture, motion compensating the particular reference picture in correspondence with the motion vectors, multiplying the motion compensated reference picture by the assigned weighting factor to form a weighted motion compensated reference picture, subtracting the weighted motion compensated reference picture from the substantially uncompressed image block, and encoding a signal indicative of the difference between the substantially uncompressed image block and the weighted motion compensated reference picture along with the corresponding index of the particular reference picture.