A Comparative Study on Single Image Dehazing using Deep Learning-Based Techniques
Deepanker Rawat, Kavinder Singh · 2023
Hazy environments significantly reduce the quality of digital images since haze varies with the scene depth. As a result, removing the haze from the original image is critical. Since deep learning techniques have been introduced, dehazing an image becomes much more efficient. This paper discusses a comparative study of different deep learning-based techniques to dehaze images. The techniques are briefly discussed along with the methodology and loss functions. The techniques have been quantitatively measured using some popular metrics like SSIM and PSNR. The paper provides a discussion of various datasets available for image dehazing. Finally, a quantitative analysis of different dehazing techniques along with the conclusion is provided.