A Comparative Study on Shadow Detection Methods Based on Features

Abstract

Shadow detection from images is an important and sensitive process before the removal process. Many methods, in the current literature, have been presented for shadow detection based on different classes of features. In this paper, we highlight some existing shadow detection methods based on different techniques and features: pixel intensity values, chromaticity, object geometrical properties, and illumination direction of the light source. We evaluate each method by introducing the features it used, summarizing algorithm work, and showing the experimental results. The paper also highlights the weakness and strengths of each method. Strengths will guide researchers to the proper use of each method on the suitable application. On the other hand, weakness will give them the opportunity for further study and improvement, or developing alternative approaches for any of those methods.