Learning a Patch Quality Comparator for Single Image Dehazing
IEEE Transactions on Image Processing
In bad weather conditions such as fog and haze, the particles present in the atmosphere scatter incident light in different directions. As a result, the image taken under these conditions suffers from reduced visibility and lack of contrast, and as a result, it appears colorless. An image dehazing method tries to recover a haze-free portrayal of the given hazy image. In this paper, we propose a method that dehazes a given image by comparing various output patches with the original hazy version and then choosing the best one. The comparison is performed by our proposed dehazed patch quality comparator based on the convolutional neural network. To select the best dehazed patch, we employ binary search. Quantitative and qualitative evaluations show that our method achieves good results in most of the cases, and are, on an average, comparable with the state-of-the-art methods.
Santra, Sanchayan; Mondal, Ranjan; and Chanda, Bhabatosh, "Learning a Patch Quality Comparator for Single Image Dehazing" (2018). Journal Articles. 1261.