A Report on Neural Network working in Words Images Understanding

The main objective of my study is to develop an automated system that could imitate human visual behavior and distinguish between non-words against pseudo-words and real-words for the scripts Kanji and German. The proposed system include development of main components as data development in appropriate form after preliminary preprocessing, feature extraction using Gabor filters, Principal Component Analysis (PCA), Neural Network (NN) training, simulation and analysis of results.