Geometric face recognition approach based on neural network

Abstract EN

In this paper, we present a technique to recognize a query person whose human face image is given w.r.t.

a specified database.

The main idea of our approach is to use the most important geometric feature positions on the half face image.

We construct vector of measures between chosen essential features points of the face such as eyes, nose, chin, mouth, face boundaries and so on, and then reduced to vector of measures.

The information taken from frontal view images, yet profile view face images are with small rotation degree to different directions.

We used the ANN technique to determine whether or not the query person is recognized.

We applied 64 human face images taken from the ORL database and 160 real human face images concerned some Iraqi persons.

The recognition results show that the recognition rate is 100 % for images taken from both databases.