Traffic Light Detector

Data preprocessing:

1. Assuming that the starting dataset has a form of videos, run

python split_video.py --pathIn *the path to a video*

to split videos into frames.

2. Rename resulting images

python rename_pics.py

3. In order to change images' contrast, run

python change_contrast.py

Making predictions:

1. Put your images to test_images folder
2. Run

python main.py

Result: