This repository is a public API for solving the problem of the final of the AIIJC competition.

The task is to create an AI for the classification of traffic signs, the ‘model’ file is a ready trained model and the prediction is done in ‘’.

The API itself is written in Python using the Flask framework

The API should receive an image of a road turn sign as input (it can consist of several signs) and send the name of the signs from left to right.

alt text

We will use our own computer as a server and ngrok as a proxy.

ngrok –

With ‘’ we start a local server

alt text

Next, open ngrok.exe and run the proxy with the command: ‘ngrok http your_port’, in our case our port is 4567, in response we will receive the URL

alt text

We will check the functionality of the API using the ‘’ file, we send a request with a photo to the URL received from ngrok.

alt text

Also we can just go to the received URL and send the file manually

alt text


View Github