Image Dehazing using Autoencoder and Vision Transformer

Baseline Dehazeformer used a vision transformer to dehaze images. I embed an autoencoder in the Dehazeformer model such that the training time is reduced by 80% with comparable SSIM and PSNR values and better segmentation results. All the procedures and improvements made in the baseline Dehazeforemer are mentioned in detail in the paper uploaded within the repository.