Performance analysis of Deep Neural Network with Hybrid Feature Extraction for traffic sign detection

  • H.Raghupathi, Dr.M.Anand ,Dr.M.JanakiRani ,Dr.K.Hemachandran

Abstract

       Driving Expert system requires the robust Traffic sign detection method to navigate the vehicle and assistance in driving.Existing method in the traffic sign detection has the lower performance in traffic sign detection due to the over fitting process. In this research, the feature extraction method such as Gabor filter, Grey Level Co-occurrence Matrix (GLCM) and Local Optimal Oriented Pattern (LOOP) are combined to analyze the image. Segmentation is based on Multi-threshold Otsu’s with cuckoo search method. Traffic sign detection is carried out based on the Support Vector Machine (SVM). Deep Neural Network (DNN) with hybrid feature extraction is applied for the traffic sign classification. The German benchmark dataset were used to evaluate the performance of proposed method. The experimental result shows that proposed hybrid feature extraction improve the efficiency of the classification.

Published
2020-05-18
How to Cite
H.Raghupathi, Dr.M.Anand ,Dr.M.JanakiRani ,Dr.K.Hemachandran. (2020). Performance analysis of Deep Neural Network with Hybrid Feature Extraction for traffic sign detection. International Journal of Control and Automation, 13(4), 378 - 390. Retrieved from https://sersc.org/journals/index.php/IJCA/article/view/16454
Section
Articles