Correct latin alphabet formation assessment using neuron values

Handwriting disability is one type of learning disability that is difficult to be detected as it may requires experts and professionals to diagnose. Owing to this matter, a computerized character recognition application is very much in need to ease the process of detecting children with learning disability based on handwriting. For any character recognition method, it's critical to extract the class correlated features from character images to maximize the mutual information. Many factors such as huge numbers of characters, noise, different fonts, various character types, and complicated document layouts pose as a challenge in developing an accurate recognition engine. This paper proposed computerized Backpropagation Neural Network (BPNN) handwriting recognition methods which assess handwriting based on the identification of errors in stroke type, sequences, and direction when forming Latin alphabets. The algorithm classified the input into three categories of stroke patterns which are straight line, complex straight line and curve. Each stroke would have its own range of BPNN neuron value. The tested handwriting values will then be compared to BPNN neuron value of reference alphabet. If the neuron value of the written alphabet is near to the neuron value of the reference alphabet, it is mean that the alphabets are well formed. The nearer the value of error rate to zero, that is mean that the alphabet written is correctly formed. The result of the testing proven that by reading the neuron values of BPNN, the computer can recognize the formation of alphabet correctly. However, it is suggested that more test data should be included to increase the accuracy of the method.