An Improved SVM-HMM Based Classifier for Online Recognition of Handwritten Chemical Symbols

In this paper, we propose an improved double-stage classfier for online recognition of handwritten chemical symbols. In the first stage, SVM based classifier is used to roughly classify chemical symbols into Non-Ring Structure(NRS) and Organic Ring Structure(ORS). Then, HMM based classifier is used for fine classification at the second stage. During the fine classification stage, we use the frequency domain feature instead of the commonly used Geometrical or Statistical feature to perform recognition task. In addition, to improve the accuracy of the ORS symbols and the speed of processing, we propose a MPSR algorithm. Finally, we achieve top-1 accuracy of 88.95% and top-3 accuracy of 98.58% on a dataset containing 9090 training samples and 3232 testing samples for 101 Chemical symbols.