This repository contains the backend, the frontend and the scientific paper in German of the project RealtimeGermanTrafficSignRecognition. The backend, developed in Python, localizes and identifies traffic signs in real time and you can choose between two different object recognition methods. OpenCV's ORB algorithm as a traditional analysis method and YOLO as a machine learning method are used for object recognition. The frontend, an Android app, uses the backend by highlighting traffic signs in a live camera preview. The scientific paper goes into more detail about the background and implementation and also compares ORB and YOLO to each other.
The project was created during our Master‘s programme at Hof University of Applied Sciences in the lecture „Data Mining and Machine Learning“.
demo-compressed.mp4
Feel free to get in contact and share your experience with RealtimeGermanTrafficSignRecognition. Bug reports are also very appreciated.