Convolutional Sparse Coding for High Dynamic Range Imaging

Abstract

Current HDR acquisition techniques are based on either (i) fusing multibracketed, low dynamic range (LDR) images, (ii) modifying existing hardware and capturing different exposures simultaneously with multiple sensors, or (iii) reconstructing a single image with spatially-varying pixel exposures. In this paper, we propose a novel algorithm to recover high-quality HDRI images from a single, coded exposure. The proposed reconstruction method builds on recently-introduced ideas of convolutional sparse coding (CSC); this paper demonstrates how to make CSC practical for HDR imaging. We demonstrate that the proposed algorithm achieves higher-quality reconstructions than alternative methods, we evaluate optical coding schemes, analyze algorithmic parameters, and build a prototype coded HDR camera that demonstrates the utility of convolutional sparse HDRI coding with a custom hardware platform.

References in corpus (2)

• Optimization Methods for Convolutional Sparse Coding
• Convolutional Matching Pursuit and Dictionary Training

Cited by in corpus (16)

• HDR image reconstruction from a single exposure using deep CNNs
• Convolutional Dictionary Learning: A Comparative Review and New Algorithms
• Convolutional Dictionary Learning: Acceleration and Convergence
• Deep BCD-Net Using Identical Encoding-Decoding CNN Structures for Iterative Image Recovery
• Attention-Guided Progressive Neural Texture Fusion for High Dynamic Range Image Restoration
• Scalable Online Convolutional Sparse Coding
• Deep Joint Demosaicing and High Dynamic Range Imaging within a Single Shot
• HDR Imaging with Quanta Image Sensors: Theoretical Limits and Optimal Reconstruction
• HDR Video Reconstruction with Tri-Exposure Quad-Bayer Sensors
• ASP Vision: Optically Computing the First Layer of Convolutional Neural Networks using Angle Sensitive Pixels
• Convolutional sparse coding for capturing high speed video content
• Adversarial Defense by Stratified Convolutional Sparse Coding
• HDR Denoising and Deblurring by Learning Spatio-temporal Distortion Models
• Convolutional Dictionary Regularizers for Tomographic Inversion
• High Dynamic Range Imaging via Visual Attention Modules
• Deep Snapshot HDR Imaging Using Multi-Exposure Color Filter Array