Dilshani Karunarathna
Traffic Sign Net
Traffic SignNet is a deep learning-based traffic sign recognition system developed to
automatically detect, localize, and classify Sri Lankan traffic signboards from road images.
Road traffic signs play a vital role in ensuring road safety by providing drivers with essential
regulatory, warning, and informational guidance. Manual recognition of traffic signs is not
feasible for intelligent transportation systems, creating the need for an automated solution
capable of accurately identifying and classifying traffic signs under varying road conditions.
The project involved training a Convolutional Neural Network (CNN) using a dataset of Sri Lankan
traffic sign images to recognize different traffic sign categories and accurately localize
signboards within input images. The developed model demonstrates the application of deep
learning and computer vision techniques to improve traffic sign recognition for intelligent
transportation systems.
Key Responsibilities
* Collected and preprocessed a dataset of Sri Lankan traffic sign images.
* Developed and trained a Convolutional Neural Network (CNN) for traffic sign classification.
* Implemented image preprocessing and data augmentation techniques to improve model performance.
* Evaluated and optimized the model using training and test datasets.
* Developed the traffic sign localization and recognition pipeline using computer vision
techniques.
* Tested and validated the system under different traffic sign scenarios.
Technologies
Python, TensorFlow, Keras, OpenCV, NumPy, Pandas, Matplotlib, Jupyter Notebook
Team
Dilshani Karunarathna
Suneth Samarasinghe
Pubudu Premathilake
Wishma Herath