Writer Identification and Verification from Intra-variable Individual Handwriting

The handwriting of an individual may vary excessively with many factors such as mood, time, space, writing speed, writing medium, utensils etc. Therefore, it becomes more challenging to perform automated writer verification/ identification on a particular set of handwritten patterns (e.g. speedy handwriting) of a person, especially when the system is trained using a different set of writing patterns (e.g. normal/medium speed) of that same person. However, it would be interesting to experimentally analyze if there exists any implicit characteristic of individuality which is insensitive to high intra-variable handwriting. In this paper, we work on writer identification/verification from offline Bengali handwriting of high intra-variability. To this end, we use two separate models for the writer identification/verification task: (a) hand-crafted features with an SVM model and (b) auto-derived features with a recurrent neural network. For experimentation, we have generated a handwriting database from 100 writers and have obtained some interesting results on training-testing with different writing speeds. Keywords—BLSTM, CNN, RNN, Speedy handwriting, SVM, Writer identification.