The goal of my AI project was to make an AI that could receive an F1 image from the 2023 season and detect it in the image as well as try to detect what team it belongs to. This model was trained on a custom-built dataset. The training took place on a Google Collab in Python with Roboflow and YoloV8.

Exemple on video : https://imgur.com/1euw5N8

Training of the AI: https://colab.research.google.com/drive/14Unf8PF688Z9TdGi0pA7UCUROOM7rMvf?usp=sharing

The online version that you can try: https://universe.roboflow.com/f1detectionaiclass/f1-car-2023/model/7

Test the AI from a VPS

It is now possible to test the AI on a picture poster directly on a VPS. The backend is a Python script using the Roboflow API and running on a FLASK server. The flask server is contacted in Ajax by an HTML/js webpage.

Selected projet for a stand at an applied science competition (Fi3E):

Présentation d’une affiche au forum innovation

VERSION FRANÇAISE

Détection de formule 1 saison 2023 par l’intelligence artificielle