Handwriting Analysis Using Machine Learning

Abstract

The culture of humans has modified since digital age where everything may be handled with a tip of the finger, however all those luxuries might return at a value of security or fraud where masking one's identity with a fake one is possible that on the opposite hand isn't possible during a case with handwriting. Handwriting is exclusive to each person like a fingerprint is exclusive to each person. Someone can imitate another person's handwriting for less than a few words creating it unique. Handwriting tells about the character of the person as writing is coupled with brain and it subconsciously leaves a path concerning the temperament attribute like openness, extraversion, consciousness etc., which might be detected. Several forms of handwriting styles are taken into thought like size of word, connecting strokes, left or right slant of the sentence, word space, latter space etc. The complete system assess the handwriting samples based on the above-mentioned handwriting styles and it is divided into four sections with the primary module being the input where the image of handwritten text is taken from the user that is followed by image pre-processing that removes noise and sharpens the difference of the image for better results, that is then passed to the Convolutional Neural Network (CNN) that analyses the input image with the CNN model which is created by performing CNN on the training dataset and classifies the input image accordingly and the last module is the output where the specified images from the earlier module is used to find out the percentage of various types of traits present in the handwriting sample of the subject.