Human behavior recognition based on hand written cursives by SVM classifier
Syeda Asra, D C Shubhangi · 2017 International Conference on Electrical, Electronics, Communication, Computer, and Optimization Techniques (ICEECCOT) · 2017
The human behavior recognition is a significant task of discovery and identification of behaviors through an input hand written image and changing that information to the corresponding machine in an editable form. This work is classified into two phases known as testing and training. In this work the hand written image is selected as input image containing a word with cursive O. In the pre-processing step of the testing phase, the image is resized, and the color conversion is also applied. The interested region is also for segmenting the middle loop cursive. The segmented middle loop is then passed to the Freeman Chain Code, and the zoning feature is extracting systems and then extracts the feature. Later in training part, the preceding trained samples of hand written images are then stored in the knowledge base. The hand writing style of cursive O is identified, and thus it defines the behavior of a person. The feature extracted system results are then compared with the SVM trained results stored in knowledge based using SVM Classifier.