Int J Performability Eng ›› 2019, Vol. 15 ›› Issue (6): 1538-1547.doi: 10.23940/ijpe.19.06.p5.15381547

Previous Articles     Next Articles

Method based on Separation Confidence Computation and Scale Synthesis Optimization for Real-Time Target Detection in Streetscape Videos

Jianmin Liua,b,*, Minhua Yangb, and Jianmei Tana   

  1. a School of Information and Statistics, Guangxi University of Finance and Economics, Nanning, 530003, China
    b School of Geosciences and Info-Physics, Central South University, Changsha, 410000, China
  • Submitted on ;
  • Contact: * E-mail address:
  • About author:Jianmin Liu received his B.S. degree from the University of Hunan, his M.S. degree from the University of Xiamen, and his Ph.D. from Central South University. His research interests include data mining and machine learning;Minhua Yang received his Ph.D. from China Agricultural University. He is currently a professor at Central South University;Jianmei Tan received her M.S. degree from the University of Technology of Changsha.
  • Supported by:
    This work was supported by the Ph.D. Research Foundation of Guangxi University of Finance and Economics, Key Research Projects of Hunan Provincial Department of Education (No. 17A108), and Guangxi Natural Science Foundation.

Abstract: This study proposes a method for the real-time detection and recognition of targets in streetscape videos. The proposed method is based on separation confidence computation and scale synthesis optimization. First, on the basis of generalization in transfer learning, we combine a fine-tuning method suitable for non-convex optimization and adaptive moment estimation in high-dimensional space. Then, we dynamically adjust the learning rates of parameters on the basis of first and second gradient moment estimations. We establish the framework and implementation steps of the proposed method by organically combining regular term super-parameter generalization and hard-example mining technology. We use the proposed method to detect and recognize targets in streetscape videos with high frame rates and high definition. Furthermore, we experimentally demonstrate that the accuracy and robustness of our proposed method are superior to those of conventional methods.

Key words: object detection, separation confidence computation, scale synthesis optimization, transfer learning, streetscape videos