Myanmar Character Identification of Handwriting Between Exhibit and Specimen

In this paper we present Myanmar character identification of handwriting between exhibits and specimen of Myanmar handwriting documents. This is also a method to identify the writer of Myanmar handwriting documents. Many methods have been reported for handwriting-based writer identification. Most such techniques assume that the written text is fixed. There are many methods for writer identification. In our method, we take the handwriting as an image containing some special individual character features, and writer identification is regarded as individual character identification. We apply the Fast Fourier Transform (FFT) method to extract features for one character. In individual character, there are character features mingled with noises. We use median filter algorithm to remove noises in this individual character features. All features were appropriately binarized so that binary feature vectors of constant lengths could be formed. We also evaluate a Weighted Euclidean Distance (WED) to compare training character features for fulfill identification task. The result of this paper will confirm whether the handwriting of the specimen is the true writer of the exhibit. The current application domain of the framework is writer identification and handwriting examination as frequently used in crime investigation and prosecution. The method is tested on 40 writers and proves to give interesting result. The identification correct rate is 97.5% in our experiments.