Lossless Pixel Data Compression via DPCM and Huffman Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current camera and smartphone buses are approaching bandwidth limits due to increasing image resolution and processing demands, and while widening and speeding up buses can help, it increases costs and complexity, with existing lossy compression methods offering limited advantages over lossless compression and complicating parallel processing.
Innovation Solution
A system and method for lossless pixel data compression using differential pulse code modulation (DPCM) encoding and Huffman encoding with a tiered technique, which compresses adjacent pixel values based on predicted values and employs restart intervals and out-of-order memory packing to reduce dictionary size and optimize data transfer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If lossy compression methods are used, then compression efficiency is improved, but image quality is degraded and error feedback complexity is introduced
Solution Approach 1:
The patent segments the compression process into two independent stages: DPCM differential compression and Huffman entropy encoding. This segmentation allows lossless compression through hierarchical processing, where the first stage compresses pixel differences and the second stage compresses the results using a dictionary that is updated but never deleted, eliminating the need for error feedback while maintaining compression efficiency.
Solution Approach 2:
The patent applies preliminary action by performing DPCM differential compression before Huffman encoding. This preliminary compression of pixel differences reduces the data volume and creates a more compressible representation, enabling efficient lossless compression without requiring error feedback mechanisms that would compromise image quality.
2Productivity
If bus bandwidth is increased to accommodate higher resolution and processing demands, then image transfer capability is improved, but device complexity and cost increase
Solution Approach 1:
The patent applies preliminary compression action at the image capture stage using DPCM and Huffman encoding, reducing the data volume before transfer. This preliminary action decreases the bandwidth requirements for the bus, allowing existing bus infrastructure to handle high-resolution images without increasing complexity or cost.
Solution Approach 2:
The patent changes the data representation parameters through differential compression and entropy encoding, transforming the pixel data into a more compact form. This parameter change reduces the effective bandwidth requirement, enabling high-resolution image transfer over existing bus infrastructure without increasing device complexity.
3Productivity
If multiple images are captured and processed in rapid succession, then image processing capability is improved, but bus bandwidth is further taxed
Solution Approach 1:
The patent applies preliminary compression to each captured image using DPCM and Huffman encoding before transfer to memory. This preliminary action reduces the data volume for each image, allowing multiple high-resolution images to be processed in rapid succession without overwhelming the bus bandwidth.
Solution Approach 2:
The patent enables continuous compression and transfer operations through the efficient two-stage encoding process. The DPCM and Huffman encoding operate continuously on captured images, maintaining high processing throughput while keeping bus bandwidth demands manageable through effective data compression.
Data Source
AI summary
A system for, and method of, pixel data compression and a smartphone incorporating the system or the method. In one embodiment, the system includes: (1) a differential pulse code modulation encoder operable differentially to compress the two pixel values losslessly to yield two losslessly compressed pixel values and (2) an entropy encoder coupled to the differential pulse code modulation encoder and configured to receive and entropy-encode the losslessly compressed pixel values using a tiered technique to yield entropy-encoded, losslessly compressed pixel values. values using a tiered technique to yield Huffman-encoded, losslessly compressed pixel values.