Interlaced Video Motion Vector Prediction Using Field Polarity Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression techniques, such as those in early versions of Windows Media Video (WMV9) and international standards like H.262, fail to adequately predict motion vectors for macroblocks of interlaced video frames, leading to inefficiencies in bit rate reduction and quality preservation.
Innovation Solution
The development of techniques and tools for computing motion vector predictors for interlaced video frames, which involve selecting between same and opposite polarity predictors based on neighborhood motion vector counts and using scaling operations to derive predictor candidates, allowing for more accurate motion vector prediction and efficient encoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional motion vector prediction techniques are used for interlaced video frames, then encoding complexity is reduced, but prediction accuracy deteriorates leading to higher bit rates
Solution Approach 1:
The patent segments the motion vector prediction process into distinct phases: generating same-polarity predictors from same-field neighboring blocks, generating opposite-polarity predictors from opposite-field neighboring blocks through scaling operations, and selectively combining them based on field polarity. This segmentation allows each predictor type to be optimized independently while maintaining manageable encoding complexity.
Solution Approach 2:
The patent changes the parameter of motion vector scaling by applying field-based scaling factors to generate opposite-polarity predictors from same-polarity source vectors. This parameter transformation enables accurate prediction across field boundaries while adapting to the specific temporal and spatial characteristics of interlaced video, thereby improving prediction accuracy without proportionally increasing complexity.
2Loss of information
If more accurate motion vector prediction is implemented, then bit rate is reduced, but encoding complexity increases
Solution Approach 1:
The patent performs preliminary generation of both same-polarity and opposite-polarity motion vector predictors before the actual motion compensation step. By pre-computing these predictors and storing them for selective use, the encoding process avoids redundant calculations during runtime, thereby reducing the bit rate through better prediction while keeping the increase in encoding complexity manageable through advance preparation.
Solution Approach 2:
The patent creates copied and scaled versions of motion vectors from neighboring blocks to generate predictor candidates. Instead of deriving all predictors from scratch, it copies existing motion vector data and applies scaling transformations, which reduces computational overhead while improving prediction accuracy through multiple candidate generators.
Data Source
AI summary
Techniques and tools for encoding and decoding predicted images in interlaced video are described. For example, a video encoder or decoder computes a motion vector predictor for a motion vector for a portion (e.g., a block or macroblock) of an interlaced P-field, including selecting between using a same polarity or opposite polarity motion vector predictor for the portion. The encoder/decoder processes the motion vector based at least in part on the motion vector predictor computed for the motion vector. The processing can comprise computing a motion vector differential between the motion vector and the motion vector predictor during encoding and reconstructing the motion vector from a motion vector differential and the motion vector predictor during decoding. The selecting can be based at least in part on a count of opposite polarity motion vectors for a neighborhood around the portion and/or a count of same polarity motion vectors.