Digit Reconition using a Multi-Layer Perceptron

This project was a done as an exercise in Machine Learning methods. The goal was to design a Multi-Layer Perceptron to use for handwritten digit classification.


This report outlines the findings of using Multilayer Perceptron methods to classify images with respect to a database of handwritten digits 0 through 9. It is found that MLPs are very effective on classifying these complex images. The results of the classification effort is also compared to a few other machine learning techniques to better illustrate the effectiveness of the MLP method.