NAVIGATR: DETECTING AND RECOGNIZING TURKISH TRAFFIC SIGNS USING A NEW DATASET BASED ON DEEP LEARNING

Abstract

While cars are becoming smarter than ever with built-in sensing technologies, thanks to the spreading availability of low-cost wearable devices, millions of cars in traffic lack such technologies. However, detecting and recognizing traffic signs is essential in ensuring the safety of pedestrians and drivers. To provide this safety, we conducted a study first to prepare a dataset using collected data in different weather conditions. Then, we used TensorFlow’s Object Detection API to detect and recognize traffic signs in Turkey. Initially, we collected over 5000 pieces of data for training. We labeled the data in the dataset using a web-based helper application and selected a suitable deep-learning model. After the training process, we evaluated the results of the models and assessed the quality of our prepared dataset. After training the model, we imported it into an Android application that we developed. This application helps navigate drivers by providing information about the signs in front of their cars using text-to-speech technology.

Keywords

• TensorFlow Object Detection API
• TTSD
• traffic sign detection.

References

1. [1] labelimg. https://github.com/tzutalin/labelImg. [Accessed 01-May-2023].
2. [2] Tuik. https://data.tuik.gov.tr/Bulten/Index?p=Road-Traffic-Accident-Statistics-2020-37436. [Accessed 01-May-2023].
3. [3] CDC. Road Traffic Injuries and Deaths—A Global Problem — cdc.gov. https://www.cdc.gov/injury/features/global-road-safety/index.html. [Accessed 01-May-2023].
4. [4] Sebastian Houben, Johannes Stallkamp, Jan Salmen, Marc Schlipsing, and Christian Igel. Detection of traffic signs in real-world images: The german traffic sign detection benchmark. In The 2013 international joint conference on neural networks (IJCNN), pages 1–8. Ieee, 2013.
5. [5] Irfan Kilic and Galip Aydin. Traffic sign detection and recognition using tensorflow’s object detection api with a new benchmark dataset. In 2020 international conference on electrical engineering (ICEE), pages 1–5. IEEE, 2020.
6. [6] Wei Liu, Dragomir Anguelov, Dumitru Erhan, Christian Szegedy, Scott Reed, Cheng-Yang Fu, and Alexander C Berg. Ssd: Single shot multibox detector. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14, pages 21–37. Springer, 2016.
7. [7] Citlalli Gamez Serna and Yassine Ruichek. Classification of traffic signs: The european dataset. IEEE Access, 6:78136–78148, 2018.
8. [8] Alexander Shustanov and Pavel Yakimov. Cnn design for real-time traffic sign recognition. Procedia engineering, 201:718–725, 2017.

9. [9] Johannes Stallkamp, Marc Schlipsing, Jan Salmen, and Christian Igel. Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognition. Neural networks, 32:323–332, 2012.
10. [10] Marco Magdy William, Pavly Salah Zaki, Bolis Karam Soliman, Kerolos Gamal Alexsan, Maher Mansour, Magdy El-Moursy, and Kerolos Khalil. Traffic signs detection and recognition system using deep learning. In 2019 Ninth international conference on intelligent computing and information systems (ICICIS), pages 160–166. IEEE, 2019.
11. [11] Jianming Zhang, Manting Huang, Xiaokang Jin, and Xudong Li. A real-time chinese traffic sign detection algorithm based on modified yolov2. Algorithms, 10(4):127, 2017.