Introduction:
Human face detection is a challenging but interesting problem since the appearance of human faces on digital (still) images strongly depends on the person's emotional state, gender, age, ethnic group and even the quality of the digital devices. The aim of this work is to implement an effective system to locate upright frontal faces on monochromatic images with use of a neural network-based classifier.
Short description of the system:
The idea of using neural networks in the task of human face detection is derived from the observation that human faces on digital images stand out for a class of well structural objects. There is a number of works applying expert’s knowledge on some characteristics of human faces (in the form of rules) to locate faces. The drawback of these methods is the need to express expert’s knowledge in an effective way (whatever this term means) and to choose the characteristics which are really important for the detecting process. Neural networks with its ability of learning can be used to extract these characteristics from a set of examples (in this case there are two sets: a set of face patterns -commonly mentioned as the set of positive samples - and the set of nonface patterns – also called negative samples). After the process of learning the structural properties of human faces should be hidden in neural networks in the form of the weights if its connections.
