• B.P. Santosh Kumar
• Shafiulla Basha Shaik
• S. Lavanya
• A. Sai Kiran Reddy
• B. Swetha , E.V.S. Srikanth Reddy , A.Rehan Babu

Haze is a common atmospheric phenomenon that significantly degrades the quality of images, affecting various computer vision applications. In this project, we propose a Pix2Pix Generative Adversarial Network (GAN) based approach implemented in PyTorch to address the task of image dehazing. The Pix2Pix GAN learns a mapping from hazy images to corresponding haze-free representations by leveraging a paired dataset of hazy and clear images. Our method employs an adversarial loss along with a pixel-wise loss function to train the generator network, enabling it to effectively learn the complex relationship between hazy and dehazed images. We evaluate the performance of our model metrics such as generator loss, discriminator loss and L1 loss during both training and validation. Additionally, we present qualitative results demonstrating the capability of our approach to produce visually pleasing dehazed images with improved clarity and detail. The experimental results showcase the effectiveness and robustness of our Pix2Pix GAN-based approach for hazy image dehazing.

A NOVEL ON END TO END SINGLE IMAGE DEHAZING USING PIX2PIX GAN

"A NOVEL ON END TO END SINGLE IMAGE DEHAZING USING PIX2PIX GAN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 4, page no.o349-o351, April-2024, Available :http://www.jetir.org/papers/JETIR2404F51.pdf

"A NOVEL ON END TO END SINGLE IMAGE DEHAZING USING PIX2PIX GAN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 4, page no. ppo349-o351, April-2024, Available at : http://www.jetir.org/papers/JETIR2404F51.pdf