Interlaced B-Frame Direct Motion Vector Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression techniques, such as those in Windows Media Video and international standards like H.262, face challenges in accurately predicting motion vectors for macroblocks of interlaced video frames, leading to inefficiencies in bit rate and distortion performance, especially in devices with limited CPU resources.
Innovation Solution
The implementation of techniques for encoding and decoding bi-directionally predicted interlaced video frames, including computing direct mode motion vectors by selecting representative motion vectors from co-located macroblocks and using 4MV coding, along with prediction mode switching and motion vector buffering, to improve motion compensation and reduce computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional motion vector prediction techniques are used for interlaced video frames, then implementation is simpler, but rate/distortion performance deteriorates due to inaccurate motion vector prediction
Solution Approach 1:
The patent segments the motion vector prediction process into distinct modes (direct mode with representative motion vector selection, 4MV coding mode, and prediction mode switching). This segmentation allows the system to choose appropriate complexity levels for different macroblock types and motion characteristics, improving overall prediction accuracy while managing computational load through selective application of complex algorithms only where needed.
Solution Approach 2:
The patent implements dynamic prediction mode switching that adapts the motion vector prediction approach based on local video content characteristics. The system dynamically selects between direct mode, 4MV coding mode, and other prediction modes on a per-macroblock basis, allowing computational complexity to vary dynamically according to actual prediction needs rather than applying a fixed complex algorithm throughout.
2Measurement precision
If direct mode motion vectors are computed by selecting representative motion vectors from co-located macroblocks, then rate/distortion performance improves, but computational complexity increases
Solution Approach 1:
The patent performs preliminary action by pre-selecting and buffering representative motion vectors from co-located macroblocks in reference frames before the actual prediction process. This preliminary selection organizes the data in advance, allowing the direct mode prediction to proceed more efficiently by simply retrieving pre-processed motion vector candidates rather than computing them on-the-fly during decoding.
Solution Approach 2:
The patent introduces an intermediary mechanism in the form of motion vector buffers that store representative motion vectors from co-located macroblocks. This intermediary structure acts as a mediator between reference frames and current frame prediction, providing pre-organized motion vector candidates that reduce the computational burden of direct mode prediction while maintaining high prediction accuracy.
3Productivity
If 4MV coding and prediction mode switching are implemented, then bit rate efficiency improves, but device complexity increases
Solution Approach 1:
The patent applies local quality by implementing 4MV coding and prediction mode switching at the macroblock level rather than uniformly across the entire frame. Different macroblocks can use different prediction modes (1MV, 4MV, direct mode, etc.) based on their local motion characteristics, allowing high bit rate efficiency where complex prediction is beneficial while maintaining lower complexity in regions where simpler prediction suffices.
Data Source
AI summary
For interlaced B-frames, an encoder/decoder computes direct mode motion vectors for a current macroblock by selecting at most one representative motion vector for each of the top and bottom fields of the co-located macroblock of the previously decoded, temporally subsequent anchor. For example, the selecting is performed based at least in part on the mode of coding the current interlaced B-frame's macroblock (e.g., 1MV mode, 2 Field MV mode, etc.). For interlaced B-fields, an encoder/decoder selects direct mode motion vectors using logic that favors the dominant polarity if the corresponding macroblock in the corresponding field of the next anchor picture was coded using four motion vectors. For example, if the corresponding macroblock's same polarity motion vectors outnumber its opposite polarity motion vectors, the encoder/decoder calculates the median of the same polarity motion vectors to obtain a motion vector for deriving direct mode motion vectors.