This paper deals with recognition of image using neural network. Since today in modern life all of us rely on internet for most of the transfer of data and for grasping up of knowledge, so the security of using internet for data transfer of data has been more important, protection technique like encryption and authentication are been used to protect data while transmitting through internet, even though the security is insufficient so the increasedsecurity on internet is done by using face recognition using neural network, even though his approach is costlier it can provide high advantage of tight security.

It is used in identifying people in real time. By combining local image samplings, the self- organizing map neural network and the convectional neural network, the self-organizing map provides the quantization ofimage samples into a topological space where inputs are in the original space and also in the output space, thereby providing dimensionality reduction and invariance to minor changes in the image sample. These software features are implemented using MATLAB v 6.5.

By implementing the face recognition system the business transaction via internet can be improved. The hybrid neural network solution which compare and recognizes a person with a large database of faces, the neural system give most likely face of people from the database often only one image is available for a person, the next stage of the system is to train the images and stored to the database. Finally the authorized person’s face is used for security monitoring software. Different representation of faces like multidimensional, complex and visual stimuli developing a computational model for face recognition is very difficult. Faces has distinguishable landmark i.e. peaks and valleys on the human faces, There are almost 80 peaks and valleys in human face. Some of them are Distance between eyes, width of nose, depth of eye, jaw line, chin, etc.

Download Artificial Intelligence for Face Recognition Project Report.