Recently, we developed a database of online handwritten isolated Bangla numerals. Samples of this database have been collected using WACOM Intuous 2 tablet. The maximum sampling rate is 200 points per second. No restriction was imposed on the writers except for specifying rectangular regions for writing each numeral sample. Each person was asked to write samples at two to five different points of time. Each sample is stored as a text file consisting of the (x, y) coordinates and pressure among a few other information of the pen tip along its trajectory. The present database consists of 6000 samples of online handwritten Bangla numerals. In each class there are 600 samples. The whole set of samples is divided randomly into training and test sets consisting of 4000 and 2000 samples respectively. A few samples from this database are shown in the following Figure.
Ref:- S. K. Parui, U. Bhattacharya, B. Shaw, K. Guin, A Hidden Markov Models for Recognition of Online Handwritten Bangla Numerals, Proceedings of the 41st National Annual Convention (CSI 2006) of Computer Society of India, D. P. Mukherjee & D. Jana (Eds.), Tata McGraw Hill Pub. Comp. Ltd. November 23-25, 2006, pp 27-31, Science City, Kolkata, India.