This repository is the code that participated in the traffic sign detection contest conducted by Halla University and Korea Road Traffic Authority and won the 'Excellence Prize'.
The model we trained has a total of 118 classes and has a high accuracy of 97%.