Similia Similibus Curantur: A reverse filtering approach to image dehazing, low-light image enhacement and high dynamic range image compression

Similia Similibus Curantur is a Latin phrase meaning that a disease may be cured by something that can cause similar symptoms. Extending this general principle to image processing leads to the concept of reverse image filtering. Given an image filter, reverse image filtering consists of recovering the original image from its filtered version by using only the filter itself in a black-box manner. In this short talk, I show how reverse image filtering can be used for image dehazing, low-light image enhancement and high dynamic range image compression. A key advantage of the considered reverse filtering approach lies in its simplicity: good image enhancement and restoration results can be achieved by using MATLAB scripts consisting of only few lines of code.