Event Analysis for Vehicle Classification using Fast RCNN
Khaing Suu Htet, Myint Myint Sein
Abstract
Event analysis on object detection is critical as classification of vehicles and persons in events. This system proposed vehicles classification and counting for event system. The images from event video streaming camera are first featured extracted, segmentation, model building and presented to achieve improving optimization by Deep Neural Learning Fast R-CNN method with hyper-parameter optimization using the two datasets, modified Stanford car dataset and new Myanmar cars dataset. Experimental result demonstrated that Fast RCNN and hyper-parameter classifier with created Myanmar cars dataset attained the state-of-art precision result in vehicle classification and counting on real-life event Video streaming.