3D Image Encoder Line-by-Line Lossless Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lossless compression techniques for digital images, particularly 3D images with range data, face challenges in achieving high compression ratios and real-time processing over bandwidth-limited communication links like Gigabit Ethernet, often requiring significant hardware resources and incurring latency.
Innovation Solution
A modified prediction-based encoding method that encodes pixel values independently per line, using a context of two preceding pixels for prediction and adaptively switching to run length mode for missing data, employing Simplified Adjusted Binary Coding (SABC) or Golomb-Rice Coding (GRC) with Run Length Coding (RLC), to achieve efficient lossless compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional lossless compression techniques are used for 3D images with range data, then compression is achieved, but processing speed and throughput are limited by hardware resources and latency
Solution Approach 1:
The patent divides the image processing into independent line-by-line operations, where each line is compressed separately using prediction based on previous lines. This segmentation allows parallel processing of multiple lines simultaneously, increasing throughput without requiring proportionally more hardware resources for sequential processing
Solution Approach 2:
The patent employs dynamic prediction models that adapt to the specific characteristics of range data from 3D imaging. The prediction algorithm dynamically adjusts based on the correlation between adjacent lines, optimizing compression efficiency for real-time processing while maintaining manageable hardware requirements
2Quantity of substance
If high compression ratios are achieved through sophisticated algorithms, then data reduction is improved, but processing latency increases
Solution Approach 1:
The patent performs prediction using data from previous lines before encoding the current line. This preliminary prediction step reduces the entropy of the data early in the process, allowing for more efficient subsequent encoding and reducing overall processing time while achieving high compression ratios
Solution Approach 2:
The patent maintains continuous processing by using already-compressed data from previous lines as input for predicting the current line. This continuous flow of useful action eliminates idle processing time and reduces latency while sustaining high compression efficiency throughout the entire image processing pipeline
3Productivity
If bandwidth-limited communication links are used for image transmission, then network cost is reduced, but image transmission speed is limited
Solution Approach 1:
The patent transforms the image data from its original high-bit-depth format into a compressed representation with fewer bits per pixel through prediction-based encoding. This parameter change in data representation reduces the total data volume by a factor of 2-4x while preserving all essential information, enabling faster transmission over bandwidth-limited links
Data Source
Figure 1A~1B
Figure 2~3
Figure 4
AI summary
Encoder (131; 800) and method for encoding of pixel values of a digital image comprising multiple lines of pixels to accomplish lossless compression of the digital image. For each of said multiple lines the encoder (131; 800) obtains (701) unencoded pixels values of the line. Further, for each of said multiple lines, the encoder (131; 800) determines (702), for each of one or more pixels of the line, which encoding to be used for encoding of the unencoded pixel value of the pixel (x) in said lossless compression of the digital image. The determination being based on how said unencoded pixel value relates to unencoded pixel values of other, closest neighboring pixels (N1, N2) of said line.