Skip to content

Latest commit

 

History

History
10 lines (6 loc) · 751 Bytes

README.md

File metadata and controls

10 lines (6 loc) · 751 Bytes

TWELO — Identification and Obfuscation of personal data in drone images

This repository contains the code for the graduation project for the Data and ML Bootcamp by Le Wagon.

Using drone images from the company Kestrix, we developed a computer vision model to identify and obfuscate personal data in drone images, ensuring data protection and seamless integration.

To develop the model, we used the pretrained model YOLOv8 and fine-tuned on kaggle. The final accuracy was 88%.

Because of the model we used, YOLO, which stands for You Only Look Once, we decided to call our project They Won't Even Look Once, short TWELO.

The final website can be seen here.