CSC790_DeepLearning

Traffic Signs Classifier

This project was meant for learning image classification using Convolutional Neural Networks.

Additionally, this project compares results between Adam and SGD optimizers.

Notebooks

• Traffic Signs Classifier using Adam optimizer
• Traffic Signs Classifier using SGD optimizer

• Both the notebooks use ipython to run on Jupyter
• This project is set to create notebook specific venv installing requirements during startup
• Dataset is automatically downloaded from github and unzipped in real-time

Concepts Used:

• Batch Normalization
• Data Augmentation
• Drop Out

Optimizers

• Adam
• SGD (Stochastic Gradient Descent)

Project Report

• Final Report
• Presentation

Results

• Screenshots

Notes and credits

• Notes
• Credits