A study regarding the generalization capacity of image classification by using neuroal networks in Matlab

الملخص EN

The paper performs an algorithmic and experimental study regarding the generalization capacity of the scheme based on neuronal networks for the recognition of new images of the face.

This enables both a rendering of graphic representations and the classification of image classes in Matlab.

The purpose is to describe the recognition algorithm, to project and implement an application which proposes both the graphic representation of the images used by the neuronal training algorithm but also the implementation of the perceptron neuronal algorithm and the determination of the generalization capacity of the separating hyper plane of the considered image classes.