Video Compression System Using Wavelet Sub-band Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current processor technologies are unable to efficiently handle the exponential increase in data from high-resolution image sequences due to limitations in clock cycle and memory size, leading to high latency and increased costs for storing and processing video data with increasing spatial and temporal resolutions.
Innovation Solution
A modular, low-cost, memory-efficient video compression system using multi-stage wavelet analysis and temporal signature analysis, which splits video frames into columns, reduces precision in higher frequency bands, and employs a novel method of temporal compression to enable low-latency transmission and processing, while allowing for arbitrary display scaling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional processor technologies are used to handle high-resolution image sequences, then processing capability is maintained at current levels, but the system cannot handle the exponential increase in data from increasing spatial and temporal resolutions
Solution Approach 1:
The image data is divided into multiple sub-bands through wavelet decomposition, separating the image into different frequency components. This segmentation allows selective processing and compression of different parts of the data, enabling the system to handle high-resolution images by processing smaller sub-band components rather than the entire high-resolution data at once.
Solution Approach 2:
Different precision levels are applied to different sub-bands based on their importance. Lower frequency sub-bands (which contain more visually important information) are processed with higher precision, while higher frequency sub-bands are processed with reduced precision. This local quality approach optimizes the balance between processing capability and visual quality, allowing the system to handle increased data volumes efficiently.
2Quantity of substance
If current memory sizes are used, then memory cost is controlled, but the system cannot store and process the increased resolution of image sequences
Solution Approach 1:
Wavelet decomposition divides the high-resolution image into multiple lower-resolution sub-bands. Each sub-band contains a portion of the original image information at reduced resolution, allowing the system to store and process these smaller sub-bands in available memory rather than requiring memory proportional to the full high-resolution image size.
Solution Approach 2:
The system changes the resolution parameter of image data by transforming the original high-resolution image into multiple lower-resolution sub-bands through wavelet transformation. This parameter change allows the same physical memory to accommodate much larger effective image data volumes, enabling handling of high-resolution sequences with limited memory resources.
3Measurement precision
If full precision is maintained in all frequency bands, then processing accuracy is preserved, but memory requirements and processing complexity increase dramatically
Solution Approach 1:
The system applies different precision levels to different sub-bands based on their visual importance. Lower frequency sub-bands (LL, LH, HL) that contain more visually significant information are processed with full or high precision, while the highest frequency sub-band (HH) is processed with reduced precision. This local quality approach maintains processing accuracy for important features while reducing overall system complexity and memory requirements.
Solution Approach 2:
The precision parameter is varied across different sub-bands rather than being uniform throughout the entire image processing system. This parameter change allows the system to maintain high accuracy where needed (in visually important sub-bands) while using lower precision (and thus lower complexity) in less critical areas, resolving the contradiction between accuracy and complexity.
4Reliability
If high-resolution video is transmitted without compression, then visual quality is maintained, but transmission latency and bandwidth requirements become unmanageable
Solution Approach 1:
The video data is segmented into wavelet sub-bands before transmission. This segmentation enables selective compression where less important high-frequency sub-bands can be more aggressively compressed or even discarded, while important low-frequency sub-bands are transmitted with higher fidelity. This allows reduced transmission time and latency while maintaining acceptable visual quality.
Solution Approach 2:
Different quality levels are applied to different sub-bands during transmission. Visually critical sub-bands are transmitted with higher quality and less compression, while less critical sub-bands are transmitted with lower quality and higher compression. This local quality approach reduces overall transmission time and latency while preserving visual quality in the most important image regions.
Data Source
AI summary
The present invention provides a method of and apparatus for operating upon a sequence of video frames by splitting each frame into components, and each component into a plurality of columns. The columns are operated upon in a manner that reduced edge artifacts and compresses the columns by reducing precision in certain higher frequency bands more than other lower frequency bands. The thus operated upon frames can be transmitted, received, and processed at a receiver with low latency and very low memory storage The invention further discusses a novel way of temporal compression using signatures of the sub bands generated for spatial compression. Spatial analysis using wavelets further enables the decoder to format and scale the decoded output to suit an arbitrary display screen. The method provides a practical solution to the problem of compressing, storing, or transmitting of video with ever-increasing spatial and temporal resolutions.