HDR Image Reconstruction via Feature Alignment and Down-Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating High Dynamic Range (HDR) images require high hardware processing capacity, are costly, resource-intensive, and lack universality due to the need for high-speed calculations to combine multiple exposures and achieve accurate scene illumination representation.
Innovation Solution
A method for reconstructing HDR images involves acquiring multiple original images with different exposure levels, performing feature alignment and fusion to create enhanced images through down-sampling and image enhancement processing, which reduces computational power while maintaining image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-speed calculation is used to combine multiple exposures and achieve accurate scene illumination representation, then HDR image quality is improved, but hardware processing capacity requirements increase
Solution Approach 1:
The patent segments the HDR reconstruction process into multiple stages: first performing feature alignment and fusion on down-sampled images to create enhanced images, then using these enhanced images for final HDR reconstruction. This segmentation allows the system to process images at lower resolution initially, reducing hardware requirements while maintaining final output quality.
Solution Approach 2:
The patent introduces a down-sampling dimension as an intermediate step between the input multi-exposure images and the final HDR output. By processing images at a lower resolution dimension first, the system reduces computational complexity while preserving the essential feature information needed for accurate HDR reconstruction.
2Measurement precision
If high-speed calculation is used to combine multiple exposures, then HDR image quality is improved, but resource consumption increases
Solution Approach 1:
The patent divides the computationally intensive HDR reconstruction into two phases: a feature extraction and fusion phase using down-sampled images (lower resource consumption), and a final reconstruction phase using the enhanced images. This segmentation significantly reduces overall resource consumption while maintaining HDR quality.
Solution Approach 2:
By introducing down-sampling as an intermediate dimension, the patent reduces the amount of data that requires high-speed processing. The system processes images at a coarser resolution initially, which consumes fewer computational resources, and only applies intensive processing to the enhanced features extracted at this lower resolution level.
3Productivity
If high hardware processing capacity is used for HDR generation, then calculation speed is improved, but device complexity increases
Solution Approach 1:
The patent segments the processing pipeline into a feature alignment/fusion stage using down-sampled images and a final HDR reconstruction stage. This segmentation allows standard hardware to perform the computationally intensive tasks through algorithmic optimization rather than requiring high-end hardware, thereby reducing device complexity.
Solution Approach 2:
The patent replaces the need for high-speed hardware calculation with an algorithmic approach that uses down-sampling and feature-based processing. Instead of relying on powerful hardware to process high-resolution images quickly, the system uses computational methods to reduce image resolution and extract features, substituting mechanical hardware capability with software algorithmic optimization.
Data Source
AI summary
A method for reconstructing an HDR image includes: acquiring a plurality of original images with same photographing scene and different exposure degrees; screening out a reference image from the plurality of original images, performing feature alignment processing on remaining original images according to the reference image, and obtaining displacement images of the remaining original images; determining an enhanced image according to the reference image and the displacement images of the remaining original images, wherein the enhanced image is obtained by performing image enhancement processing on a fused image after a down-sampling operation, and the fused image is obtained by performing feature fusion on the reference image and the displacement images of the remaining original images; and reconstructing the HDR images corresponding to the plurality of original images according to the enhanced images.