Steerable Pyramid-based Multi-Scale Fusion Algorithm for Single Image Dehazing

Abstract

Multi-Scale Fusion Technique employs different versions of the image to be dehazed to extract haze-relevant parameters that aid in image dehazing. The paper proposes a variant of the technique using a Steerable Pyramid has been proposed. The proposed method extracts steer-coefficient matrices from the original hazy image and incorporates them into the multi-scale fusion algorithm. The proposed technique aims to reduce the halo effect observed in image-dehazing applications and related works. Moreover, this paper aims at the quality improvement of the output image by the novel algorithm it proposes. Quantitative, as well as a visual display of results, using the I-HAZE dataset, give an accurate interpretation of the effectiveness of the proposed work.