A Review on Challenges and Applications of Digital Graphology

摘要

Handwriting is a process of using letters and symbols that follows a set of guidelines to communicate with others. It is a behavioral biometric that requires verification system to discriminate whether a signature is genuine or forged. In this paper a review of datasets, acquisition devices and preprocessing for offline and online systems followed by characters of handwriting like margins, slanting lines, upward writing, connected letters, speed and pressure points that deduce the writer's personality traits are proposed. The current state of the art has been discussed along with the accuracy. Support Vector Machines, Hidden Markov Model, Neural Networks, Dynamic Time Wrapping algorithms and Symbolic Representations are the promising approaches that has been mentioned in the later part of the paper. The final division in the paper highlights the applications of this system such as -Identifying heart and Parkinson's diseases, signature analysis and psychological assessment specifically on emotional state at critical events.