Sparse-to-Dense Image Reconstruction With Residual Error Randomization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electrical devices face challenges in efficiently generating high-quality image data while reducing development costs by utilizing a common image signal processor for both sparse and dense image data formats.
Innovation Solution
A method involving obtaining sparse image data, extracting and splitting data parts, inversely converting residual data using a compression curve, adding random values within an error range, and reconstructing dense image data to improve image quality while optimizing processor usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a common image signal processor is used to generate both sparse and dense image data, then development costs are reduced, but image quality and processing efficiency deteriorate due to data loss and jaggies
Solution Approach 1:
The patent segments the dense image data processing into two parts: original sparse image data and reconstructed residual data. The sparse image data is processed through compression and embedding, while the residual data is separately reconstructed and then combined, improving overall image quality while using a common processor
Solution Approach 2:
The patent applies parameter changes by using compression curves to transform residual data into compressed form, then inversely converting it back. This transformation process reduces data loss and eliminates jaggies while maintaining compatibility with the common image signal processor
2Productivity
If sparse image data is compressed and embedded, then data transmission efficiency is improved, but image reconstruction accuracy deteriorates due to compression loss
Solution Approach 1:
The patent uses compression curves to transform residual data into compressed form for efficient transmission, then applies inverse conversion to reconstruct the original data. This parameter transformation approach maintains reconstruction accuracy while improving transmission efficiency
Solution Approach 2:
The patent implements a feedback mechanism where the compressed data is inversely converted and compared with the original residual data. The error range analysis provides feedback to adjust the reconstruction process, ensuring high accuracy is maintained despite compression
Data Source
AI summary
A method includes obtaining an embedded sparse image data for generating a target image data from an image signal processor, wherein the embedded sparse image data includes a sparse image data including pixels which include at least first color pixels, second color pixels and third color pixels; extracting a split data including first and second data parts from the sparse image data; joining the first and second data parts to obtain a compressed data; inversely converting the compressed data based on a compression curve to obtain an inversely converted residual data; adding a random value within an error range to the inversely converted residual data to obtain a reconstructed residual data; and reconstructing a dense image data based on the reconstructed residual data and the sparse image data, wherein the dense image data includes pixels including the first color pixels.