Image Encoding via Wavelet Segmentation for Memory Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-resolution video processing in portable devices faces high memory access loads due to inefficient compression methods, particularly with wavelet multiresolution processing, which results in high latency and power consumption, and is not optimized for real-time applications.
Innovation Solution
A method involving wavelet transformation, quantization, entropy coding, and zero run length encoding to derive exponents and mantissas for efficient image compression, reducing memory requirements and latency while maintaining high video quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If wavelet multiresolution processing is used for frame buffer compression, then memory access load is reduced, but latency increases and power consumption becomes unacceptably high
Solution Approach 1:
The patent segments the wavelet transformation process into multiple passes, where each pass processes a specific resolution level. This allows the system to process only the necessary portions of the image data at each resolution level, reducing the total computational load and latency while maintaining memory compression benefits.
Solution Approach 2:
The patent performs preliminary downsampling and coarse-level wavelet transformation before detailed processing. By pre-processing the image to reduce its resolution and complexity first, the system minimizes subsequent processing requirements, thereby reducing overall latency and power consumption.
2Loss of information
If multiple read passes through the frame are used for compression, then compression quality is improved, but memory requirements increase and latency increases
Solution Approach 1:
The patent divides the compression process into multiple resolution levels, with each level processed separately. This segmentation allows the system to achieve high compression quality through multiple passes while using less memory at any given time, as each pass processes only the current resolution level's data.
Solution Approach 2:
The patent introduces a resolution level dimension to the processing approach. Instead of processing the entire high-resolution frame multiple times, the system processes different resolution levels in sequence, effectively trading temporal complexity for spatial resolution management, thereby reducing memory requirements.
3Productivity
If h.264 compression methods are used, then compression efficiency is improved, but compatibility with wavelet multiresolution processing is poor
Solution Approach 1:
The patent modifies the wavelet transformation parameters and processing stages to align with h.264 compression requirements. By adjusting the wavelet decomposition levels, block sizes, and transformation characteristics, the system achieves compatibility between wavelet multiresolution processing and h.264 entropy coding, thereby maintaining both compression efficiency and adaptability.
Data Source
AI summary
A method for processing an image, the method comprising retrieving an image, encoding the image as a string of components, deriving an exponent for each component, deriving mantissas wherein at least an approximation of each component can be derived from the exponents and mantissas, and wherein each exponent indicates the number of bits in its accompanying mantissa, compressing at least the exponents, and storing the exponents and the mantissas in a memory. There is also provided a apparatus for processing an image.