A Comparative Study of CNN, SSD and VGG16 for Robust Traffic Sign Recognition
DOI:
https://doi.org/10.65204/djes.v3i2.711Keywords:
Traffic light , Traffic jam, Deep learning, Transfer learningAbstract
These techniques do everything in their power to get it right to avoid hazardous mistakes. The downside, they can also be too slow for everyday driving, where the decisions need to happen within a blink of an eye. Focus on Speed: Such methods run very fast. The downside, they may not be consistent enough at challenging times, such as reading signs far away, under darkness, or at bad weather. This thesis overcomes these challenges through combining deep learning-a sophisticated type of AI-with an intelligent image pre-processing technique. We have chosen deep learning because it has been proven to perform very well in a wide range of analysis task, even outperforming or matching human capabilities in certain visual tasks. In our tests, we implemented different deep learning models-CNN, SSD, and VGG16-based on the strengths and weaknesses of our image dataset. Our experience showed that the simplest CNN model was not fit for this purpose. On the other hand, the VGG16 model performed better. It was capable of recognizing traffic signs reasonably well, including under complex situations such as variable weather conditions or lighting, or when the traffic signs were far away. The best performing model is VGG16 with 91% testing accuracy.
References
M. P. Philipsen, A. Møgelmose, T. B. Moeslund and M. M. Trivedi, "Vision for Looking at Traffic Lights: Issues, Survey, and Perspectives,", IEEE transactions on intelligent transportation systems, 2016, 17.7, 1800-1815.
K. He, X. Zhang, S. Ren and J. Sun, "Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification," in Proceedings of the IEEE nternational conference on computer vision, pp. 1026–1034, 2015.
G. D, L. S and Y. X, "Recognition of traffic signs by artificial neurale networks," In Proceedings of ICNN'95-International Conference on Neural Networks, 1995, Vol. 3, pp. 1444-1449.
K. N and E. L, "A real-time histography approach to road sign recognition," In Proceeding of Southwest Symposium on Image Analysis and Interpretation, 1996, pp. 95-100.
Y. N, A.-A.-N. D and M. M, "Road traffic sign detection in color images," presented at 10th IEEE Inter. Conf. on Electronics, Circuits and Systems (ICECS 2003), Sharjah, United Arab Emirates, 2003, Vol. 2, pp. 890-893.
C. F. S. C. a. P. Y. C. Fang, "A road sign recognition system based on dynamic visual model," In 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2003. Proceedings, 2003, Vol. 1, pp. I-I.
R. L. a. V. R. L. Priese, "Ideogram identification in a realtime traffic sign recognition system," In Proceedings of the Intelligent Vehicles' 95, 1995, pp. 310-314.
Y. A. a. T. Asakura, "A study on traffic sign recognition in scene image using genetic algorithms and neural networks,” In Proceedings of the 1996 IEEE IECON. 22nd International Conference on Industrial Electronics, Control, and Instrumentation, 1996, Vol. 3, pp. 1838-1843.
Y. A. a. T. Asakura, "Detection and recognition of traffic sign in scene image using genetic algorithms and neural network," In Proceedings of the 35th SICE Annual Conference. International Session Papers, 1996, pp. 1343-1348.
C. Schiekel, "A fast traffic sign recognition algorithm for gray value images," In International Conference on Computer Analysis of Images and Patterns, Berlin, Heidelberg: Springer Berlin Heidelberg 1999, pp. 588-595.
P. P. a. J. Novovicova, "Road sign classification without color information," In Proc. of the 6th Annual Conference of the Advanced School for Computing and Imaging, 2000.
L. Shangzheng, "A Traffic Sign Image Recognition and Classification Approach Based on Convolutional Neural Network," in 2019 11th International Conference on Measuring Technology and Mechatronics Automation (ICMTMA), 2019, pp. 408–411.
T. S. Tsoi and C. Wheelus, "Traffic Signal Classification with Cost-Sensitive Deep Learning Models," in 2020 IEEE International Conference on Knowledge Graph (ICKG), Aug 2020, pp. 586–592.
M. A. S. Al-Hitawi, A. L. Alzaidy, M. A. Alazaizi, and M. A. Tharthar, “Toolkit for Generating and Augmenting Hungarian Handwritten Text Recognition Dataset,” Dijlah J. Eng. Sci., vol. 3, no. 1, 2026.
P. Dhar, M. Z. Abedin and T. Biswas, "Traffic sign detection — A new approach and recognition using convolution neural network," in 2017 IEEE Region 10 Humanitarian Technology Conference (R10-HTC), 2017, pp. 416–419.
Al-Hitawi MAS and Máté GN. Enhancing Transformer-Based Language Models for Hungarian Handwritten Text Recognition [version 1;peer review:1 approved with reservations]. F1000Research 2026, 15:181 (https://doi.org/10.12688/f1000research.176408.1).
Y. Wu, Y. Liu, J. Li, H. Liu and X. Hu, "Traffic sign detection based on convolutional neural networks," in 2019 International Conference on Advances in Computing, Communication and Control (ICAC3), 2019, pp. 1–6.
