Cycle-Spinning GAN for Raindrop Removal from Images

Abstract

Weather events such as rain, snow, and fog degrade the quality of images taken under these conditions. Enhancement of such images is critical for intelligent transport and outdoor surveillance systems. Generative Adversarial Networks (GAN) based methods have been shown to be promising for enhancing these images in recent years. In this study, we adapt the cycle-spinning technique to GAN for removal of raindrops. The experimental evaluation of the proposed method shows that the performance is improved in terms of reference-based metrics (SSIM and PSNR). In addition, the approach also results in higher object detection performance in terms of mean average precision (mAP) metric when applied before the detection process.