Interacting with PC using Facial Gesture Recognition

Features Extraction Using the positions of 20 facial landmarks of each image, 19 dependent features will be exctracted by computing Euclidean distance between the position of each facial landmark and the position of tip of nose

Classification Algorithms Implemented the following two learning algorithms for recognizing the facial gestures in ONE package:

1 Multilayer Perceptron --> Back-Propagation 2 Radial-Basis Function --> Least Mean Square

The dataset is real-world data, gathered from BioID Face Database [1]. The datasetconsists of 80 gray level images (patterns) (20 images/class) with a resolution of (384*286) pixel. Each one shows the frontal view of a face of one out of 23 different test persons

Using the positions of 20 facial landmarks of each image, 19 dependent features will be exctracted by computing Euclidean distance between the position of each facial landmark and the position of tip of nose