3D Video Frame Buffer Compression via Base and Residual View Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for processing 3D video are inefficient in terms of frame buffer compression and memory resource utilization, leading to increased bandwidth and storage requirements.
Innovation Solution
A system and method that compresses 3D video by generating a base view and an enhancement view, where base view pictures not used for prediction are dropped, and residual views are generated by subtracting base view pictures from enhancement view pictures, allowing for reduced memory resource usage through selective buffering and multiplexing for transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional methods are used for processing 3D video, then video quality is maintained, but memory bandwidth and storage requirements increase significantly
Solution Approach 1:
The patent segments the 3D video into a base view and an enhancement view. The base view contains the essential video information, while the enhancement view contains additional details. By processing and transmitting only the base view and its residuals, the system reduces memory bandwidth and storage requirements while maintaining acceptable video quality through selective buffering and transmission of only necessary data.
Solution Approach 2:
The patent extracts and removes redundant information from the 3D video processing pipeline. Specifically, it drops base view pictures that are not used for prediction and transmits only the enhancement view residuals along with motion vectors. This extraction of unnecessary data significantly reduces the quantity of data requiring memory bandwidth and storage while preserving the essential video content.
2Measurement precision
If all base view pictures are buffered for prediction, then prediction accuracy improves, but memory resource consumption increases
Solution Approach 1:
The patent applies partial action by buffering only the necessary base view pictures required for prediction rather than all available pictures. It selectively buffers base view pictures that will be used for prediction of enhancement view pictures, while dropping pictures that are not needed. This partial buffering approach reduces memory resource consumption while maintaining sufficient prediction accuracy for the essential video content.
3Quantity of substance
If enhanced view pictures are transmitted without compression, then video quality is preserved, but bandwidth requirements increase
Solution Approach 1:
The patent extracts and transmits only the residual information from the enhancement view rather than the complete enhanced view pictures. By transmitting only the residuals along with motion vectors, the system significantly reduces bandwidth requirements. The receiver can reconstruct the enhancement view pictures from the base view and residuals, preserving video quality while using much less bandwidth than transmitting uncompressed enhanced views.
Data Source
AI summary
A video receiver receives a compressed 3D video comprising a base view video and a residual view video from a video transmitter. The video receiver decodes the received base view video and an enhancement view video of the received compressed 3D video into a left view video and a right view video. Base view pictures are generated selectively based on available memory resource. The residual view video is generated by subtracting base view pictures from corresponding enhancement view pictures. The received base view and residual view videos are buffered for video decoding. Pictures in the buffered residual view video are added to corresponding pictures in the buffered base view video for enhancement view decoding. The left view video and/or the right view video are generated from the resulting decoded base view and enhancement view pictures. A motion vector used for a disparity predicted macroblock is applied to adjacent macroblock pre-fetching.