Recovering the Unseen:
Benchmarking the Generalization of Enhancement Methods to Real World Data in Heavy Fog
written by: Mario Bijelic, Paula Kysela, Tobias Gruber, Werner Ritter, Klaus Dietmayer
Due to the ill-posed problem of collecting data within adverse weather scenarios, especially within fog, most approaches in the field of image de-hazing are based on synthetic datasets and standard metrics that mostly originate from general tasks as image denoising or deblurring. To be able to evaluate the performance of such a system, it is necessary to have real data and an adequate metric. We introduce a novel calibrated benchmark dataset recorded in real, well defined weather conditions. The aim is to give a possibility to test developed approaches on real fog data. Furthermore, we claim to be the first showing an investigation of heavy fog conditions up to a total degradation of the considered images. We present a newly developed metric providing more interpretable insights into the system behavior and show how it is superior to several current evaluation methods as PSNR and SSIM. For this purpose, we evaluate current state-of-the-art methods from the area of image defogging and verify the proposed dataset and our developed evaluation framework.
You should definitely visit thecvf.com and take a look on our paper.
Currently data and code are available upon request.