Feature-Centric Image Enhancement via Dehazing

Abstract

In this paper, we address the image dehazing problem through a global feature-restoration pipeline. We propose a dark channel prior-based global image dehazing algorithm which captures and restores the true features of pixels within haze-degraded regions by applying scene depth selection and adaptive filtering. Our scheme harnesses haze and depth features intuitively across a given image without the prior scene depth information. This allows our scheme to sustain a high dehazing efficiency across all image regions irrespective of the local depth variations. We prove that haze degradation is linearly correlated with scene depth and based on this nuance, propose a depth selection and cropping scheme, which guides the adaptive filter iteratively across the image. Secondly, we put forward haze relevant image features and highlight the dark-channel prior for image dehazing. We merge the dark channel prior and scene depth-cropping schemes into a unified dehazing pipeline which is capable of sustaining uniform and robust results across all image regions, in real-time. We verify the superiority of the proposed scheme in terms of speed and robustness through computer-based experiments. Finally, we present comparison results with state-of-the-art and further highlight the comparative superiority of our scheme.