An Improvement of Traffic Incident Recognition by Deep Convolutional Neural Network
Hoai Nam Vu1, Ngoc Hung Dang2
1Hoai Nam Vu, Department of Computer Science, Posts and Telecommunications Institute of Technology, Ha Noi, Viet Nam.
2Ngoc Hung Dang, Department of Computer Science, Posts and Telecommunications Institute of Technology, Ha Noi, Viet Nam.
Manuscript received on 10 November 2018 | Revised Manuscript received on 25 November 2018 | Manuscript published on 30 November 2018 | PP: 10-14 | Volume-8 Issue-1, November 2018 | Retrieval Number: A2529118118/18©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Traffic incident is the source of many problems that cause economic and human life damages, especially in developing countries. There has been a great amount of research which focuses on early warning traffic incident on the road. Although some researches are able to achieve promising results, the problem of traffic incident detection is still far from completely solved due to the difficult situations such as weather condition, a group of vehicles traveling at the same time. In this paper, we propose a method, which takes advantage of recent deep learning model in vehicle detection and recognition to detect traffic event on separate lanes. Experimental results on real-world dataset prove that the proposed method is effective in locating incidents happing while ensuring real-time scenario of the system. Index Terms:
Keyword: Deep Convolutional Neural Network, Traffic Incident, Pattern Recognition
Scope of the Article: Neural Network