Global Feature Fusion Attention Network For Single Image Dehazing

As a common low-level visual task, image dehazing is widely used in various multimedia fields. However, it faces great challenges due to its ill-posed property, especially the problem with dense haze removal. In this paper, we propose an efficient end-to-end global feature fusion attention network for single image dehazing based on an encoder-decoder structure. We design a Multi-scale Global Conte...