Face detection and emotion recognition based on 8-connected neighbors approach

Abstract EN

Face detection and emotion recognition has become an important issue in many applications in computer visions communication and automatic access control system.

This paper proposes accurate and robust detection system for face and lip.

The algorithm recognizes the face from images and ignores the rest.

It depends on color of the skin and edges of the face.

Irrelevant regions are removed from images by applying Morphological connected transformations.

After face detection the system divides the face to objects and then detects the lips and recognizes them.

The algorithm depends on the connected components analysis without boundaries.

It also adopted the 8-connected neighbors approach.

A proposal work is used to recognize some of facial features.

Additionally, the algorithm compares the detected lip with set of templates to extract the facial emotion.