A Machine Learning Technique for Identification of Personality Traits Using Handwriting

Abstract

Personality analysis using handwriting is a method of understanding the personality traits of a person depends on the way person writes. Handwriting may be unique for each person and a person’s nature, behavior, and certain psychological aspects can be inferred based on it. The field that deals with study of handwriting analysis of a person termed as Graphology. The people’s works on handwriting analysis called as Graphologist .Research has proven that we make more use of brain while writing than the speaking. After the digital age, the way of life has changed. Anything can be done with a fingertip, but all these luxuries are threatened at security or fraud costs. Handwritten script or calligraphy explores the personality of an individual. It speaks about the person’s  character and predicts an Attribute as positive, Social Maturity, balanced and shy inside calligraphy because writing is related to the brain and leaves a sample subconsciously. Several types of handwriting features are regarded as slope, baseline, top margin, word size, line spacing, word spacing, left or right or ordinary slant or irregular expression. The entire device tests the script based on the aforementioned types of calligraphy and divides the image into three modules with the main module, which is used to eliminate noise and to sharpen the picture’s contrast to enhance the performance. Extract the 7 features of each dataset picture, and then use CNN in combination with Multi-Layer Perceptron (MLP). Compared to the literature survey, the proposed framework gives a better result. The process of Graphology is a manual process of understanding the handwriting of individual person .It takes time for understanding the handwriting of the person. Moreover it is costly and depends on the understanding of the Graphologist. So to make this process of handwriting recognition easier without consuming time we will extract the important features from the dataset that will help us to identify the personality traits. For this analysis we have used the IAM dataset consist of 650 writers. The IAM dataset consist of handwriting images .Various image processing techniques has been applied on the images of A4 size paper written using English language sentences. The machine learning algorithms have been used to identify the various personality traits of the person .It includes the manual process of feature extraction. The accuracy of the machine learning algorithms vary from 70-87 % .The proposed method of handwriting based personality analysis gives 94% accuracy.