SDR to HDR Conversion via Convolutional Networks and 3D Lookup Tables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for converting Standard-Dynamic Range (SDR) images to High-Dynamic Range (HDR) images require manual toning and lack automation, which is labor-intensive and inefficient.
Innovation Solution
An image processing method utilizing a first convolutional network for feature analysis and a second convolutional neural network for refinement correction, along with 3D lookup tables, to automatically convert SDR images to HDR images, involving preprocessing, color adjustment, and refinement processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual toning is used for HDR conversion, then color quality can be adjusted, but labor cost increases and production efficiency decreases
Solution Approach 1:
The system performs automatic toning adjustment through convolutional neural networks and 3D lookup tables, enabling the HDR conversion process to self-adjust color parameters without manual intervention, thus resolving the contradiction between color quality and production efficiency
Solution Approach 2:
The patent replaces manual mechanical toning operations with automated computational systems including convolutional networks and lookup table algorithms, substituting human labor with computational processing to maintain color quality while dramatically improving productivity
2Manufacturing precision
If manual toning is used for HDR conversion, then color adjustments can be made, but labor cost increases
Solution Approach 1:
The automated system performs color adjustments independently using trained convolutional networks that automatically learn optimal toning parameters, eliminating the need for manual labor while maintaining precise color control
Solution Approach 2:
The system transforms color adjustment from manual parameter setting to automated parameter optimization by using convolutional networks to learn and apply optimal color transformation parameters automatically
3Productivity
If automated methods are used for HDR conversion, then production efficiency improves, but color quality may deteriorate
Solution Approach 1:
The patent replaces manual toning with automated convolutional neural networks that have been trained to preserve color quality, achieving both high productivity and maintained color fidelity through computational processing
Solution Approach 2:
The system incorporates feedback mechanisms where the convolutional network learns from training data and continuously optimizes color transformation parameters to maintain high color quality while operating automatically at high speed
Data Source
AI summary
The present disclosure provides an image processing method and an image processing apparatus. The image processing method includes: obtaining a to-be-converted SDR image; using a first convolutional network to perform feature analysis on the SDR image, to obtain N weights of the SDR image; where the N weights are respectively configured to characterize proportions of color information of the SDR image to color information characterized in preset N 3D lookup tables, the N 3D lookup tables are configured to characterize color information of different types; obtaining a first 3D lookup table for the SDR image according to the N weights and the N 3D lookup tables; using the first 3D lookup table to adjust the color information of the SDR image to obtain an HDR image; and using a second convolutional neural network to perform refinement correction on the HDR image to obtain an output image.