Considering StatPR

We explored statistical character recognition for our discretemathmodeling project. In statistical character recognition, a feature vector is used to classify characters. A feature vector represents in numerical form the important characteristics of an input. From the feature vector, classification is done. The feature vector length is important because it will allow us to recognize better as it increases, but makes computing slower at the same time. We wrote a program to recognize 10 Japanese handwritten characters. Then, we tested the program to see how good the recognition is. A vector of length sixty-four was found to yield a 97% recognition rate. These samples were people’s own handwritten characters. The program uses basic statistical analysis to find the mean and standard deviation of character’s vectors.