HANDWRITTEN DIGIT RECOGNITION BY PCA AND KOHONEN'S SELF-ORGANIZING MAP (SOM)

摘要

The Principal Components Analysis is a very classical method in pattern recognition. PCA reduces the sample dimension in a linear way for the best representation in lower diixiensions keeping the maximum of inertia. In recent times, a great of interest in the study of artificial neural networks apt to computing a principal component extraction has been observed. The present work is devoted to the description and performance analysis (by means of computer simulations) of some neural networks of such a kind. It has been used unsupervised neural network with Hebbian law of synaptic weights' adjustment to extract the features of the pattern. The actual number of eigenvalues has been determined by simulating the outputs of PCA.