Feature extraction using marker based watershed segmentation on the human face

In this paper, we propose to segment the human face using marker based watershed segmentation, an improved version of watershed segmentation. This algorithm uses the concept of markers as a pre requisite to determine the segments. The proposed method extracts facial features, most importantly the eyes and the lips. The successful feature extraction of the human face can contribute in a significant way to a face recognition algorithm based on the extracted features. The algorithm has been tested on colored face images of various races, obtained from standard databases and produced positive results.