Image Enhancement Network Trained by Using HDR images

TL;DR

This paper introduces a new image enhancement network trained with HDR images that outperforms existing methods in quality and can restore lost pixel values, using a simpler architecture.

Contribution

The paper presents a novel CNN-based image enhancement network trained with HDR images that achieves higher performance and restores pixel values lost due to clipping.

Findings

Outperforms conventional enhancement methods in quality metrics

Restores pixel values lost by clipping and quantizing

Uses a simpler network architecture than existing CNN methods

Abstract

In this paper, a novel image enhancement network is proposed, where HDR images are used for generating training data for our network. Most of conventional image enhancement methods, including Retinex based methods, do not take into account restoring lost pixel values caused by clipping and quantizing. In addition, recently proposed CNN based methods still have a limited scope of application or a limited performance, due to network architectures. In contrast, the proposed method have a higher performance and a simpler network architecture than existing CNN based methods. Moreover, the proposed method enables us to restore lost pixel values. Experimental results show that the proposed method can provides higher-quality images than conventional image enhancement methods including a CNN based method, in terms of TMQI and NIQE.