Scalable Tile Processing for Image Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems employing complex algorithms like JPEG 2000 for compression and decompression are computationally intensive, leading to increased time requirements for processing large images in real-time applications.
Innovation Solution
A scalable tile processing system that partitions image data into tiles, performs wavelet transformation, quantization, and encoding using integer operations, generating encoded bit streams for efficient compression and decompression, while reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If floating point-based wavelet transform and quantization are used to achieve enhanced compression efficiency, then image quality and compression ratio are improved, but computational complexity increases and processing time extends
Solution Approach 1:
The patent replaces floating point arithmetic operations with integer arithmetic operations in the wavelet transform and quantization processes. Specifically, it uses integer lifting coefficients instead of floating point coefficients, and implements quantization using integer multiplication and division operations. This substitution maintains compression efficiency while dramatically reducing computational complexity and processing time.
Solution Approach 2:
The patent changes the numerical parameter representation from floating point to integer format. It introduces integer lifting coefficients (a, b, c, d) that replace floating point wavelet coefficients, and uses integer base quantization parameters. This parameter transformation enables the system to achieve the same compression performance with significantly reduced computational burden.
2Manufacturing precision
If complex JPEG 2000 algorithms are used to achieve enhanced compression efficiency, then compression ratio is improved, but processing speed decreases
Solution Approach 1:
The patent substitutes computationally intensive floating point operations with efficient integer operations throughout the JPEG 2000 compression pipeline. The integer lifting scheme replaces floating point wavelet transforms, and integer quantization replaces floating point quantization. This substitution maintains the enhanced compression efficiency of JPEG 2000 while enabling real-time processing speeds.
Solution Approach 2:
The patent divides the image into tiles and processes each tile independently using the integer-based wavelet transform. This segmentation enables parallel processing of multiple tiles, further improving processing speed while maintaining compression efficiency. Each tile is transformed, quantized, and encoded separately using the efficient integer arithmetic operations.
Data Source
AI summary
A system for processing an image includes a an image data input port, a compressed image data output port or a compressed image data storage node, a non-transitory memory component storing a set of executable instructions, and a scalable tile processing device. The executable instructions cause the system to receive image data, partition the image data into tiles, transmit a tile to the scalable tile processing device, receive an encoded bit stream corresponding to the transmitted tile from the tile processing device, and output compressed image data including the encoded bit stream. The scalable tile processing device receives the tile including tile image data, wavelet transforms, quantizes, segments, and encodes the tile image data to generate a plurality of encoded code blocks, and transmits an encoded bit stream including the plurality of encoded code blocks to the system.