Die nachfolgende Aufgabenstellung ist in englischer Sprache verfasst. Die Bearbeitung der Themen kann, je nach Wunsch des Studenten, wahlweise in Englisch oder Deutsch erfolgen.

Description

Skin detection is a common problem in computer vision. The so called pixel-bases skin detectors decide if a certain pixel in an image belongs to skin based on its RGB or HSV values,   learned from a database of labeled images. Most of these approaches do not consider, however, the situations that appear in real applications, like real-time requirements and specially, changing light conditions. 

This bachelor thesis will focus around the development of a skin detection algorithm that takes into account these factors, and not only works on operates on static images but on real-time data obtained from the robot. The algorithm should be able to operate in real time and in multiple type of environments, adjusting itself to the environment conditions using learning algorithms.

Also the tracking of skin "blobs" and the integration of the algorithm with face recognition to enable human detection will be considered.

 

Literature

A Survey on Pixel-Based Skin Color Detection Techniques - Vladimir Vezhnevets

 

If you are interested:

Dipl-Ing. Roderick de Nijs

rsdenijs@lsr.ei.tum.de