Hybrid Loss for Learning Single-Image-based HDR Reconstruction
arXiv:1812.07134
Abstract
This paper tackles high-dynamic-range (HDR) image reconstruction given only a single low-dynamic-range (LDR) image as input. While the existing methods focus on minimizing the mean-squared-error (MSE) between the target and reconstructed images, we minimize a hybrid loss that consists of perceptual and adversarial losses in addition to HDR-reconstruction loss. The reconstruction loss instead of MSE is more suitable for HDR since it puts more weight on both over- and under- exposed areas. It makes the reconstruction faithful to the input. Perceptual loss enables the networks to utilize knowledge about objects and image structure for recovering the intensity gradients of saturated and grossly quantized areas. Adversarial loss helps to select the most plausible appearance from multiple solutions. The hybrid loss that combines all the three losses is calculated in logarithmic space of image intensity so that the outputs retain a large dynamic range and meanwhile the learning becomes tractable. Comparative experiments conducted with other state-of-the-art methods demonstrated that our method produces a leap in image quality.
20 pages, 17 figures
References in corpus (9)
- Adam: A Method for Stochastic Optimization
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- HDR image reconstruction from a single exposure using deep CNNs
- Texture Synthesis Using Convolutional Neural Networks
- Deep Chain HDRI: Reconstructing a High Dynamic Range Image from a Single Low Dynamic Range Image
- Attentive Generative Adversarial Network for Raindrop Removal from a Single Image
- Image Correction via Deep Reciprocating HDR Transformation
- Learning an Inverse Tone Mapping Network with a Generative Adversarial Regularizer
- Fully Convolutional Network with Multi-Step Reinforcement Learning for Image Processing