Traffic sign recognition model

Abstract

The application of information technology in solving problems in the field of traffic transport is a subject undergoing intense study recently. Intelligent traffic development to reduce congestion, accidents and resources is also discussed in various forums and conferences, such as the Vietnam Information and Communication Technology Conference (Vietnam ICT Summit) 2015. Road sign detection and recognition are considered as a support method in the intelligent traffic system. These kinds of systems are being developed and applied in smart automation industry in the world. In terms of theoretical research, traffic sign recognition is a topic that receives close review in decades. In Vietnam, famous work in this field that can be listed are "Real time traffic sign detection and recognition base on SVM algorithm" by Le Thanh Tam in 2009 or "Traffic sign recognition base on local features algorithm" by Nguyen Duy Khanh in 2011. However, the results reported above are not comparable. As all systems are evaluated by proprietary data, most of which are not publicly available. Therefore, we present a freely available, extensive traffic sign dataset to allow unbiased comparison of traffic sign recognition approaches. In this project, we study the topic “Traffic Sign Recognition Model” under the guidance of PhD. Ho Phuoc Tien and PhD. Huynh Huu Hung. The model includes two main parts: Recognition and Detection. Up-to-date, the project successfully builds a complete model that can detect and recognize traffic sign in images with a high performance.