AI-Based Image Detection and Processing using PyTorch, OpenCV, and YOLOv8

A flowchart showing how the program works

In this study, I developed a project to reinforce and apply the theoretical knowledge I have learnt about artificial intelligence. In my project, I wanted to develop and use my own trained model instead of pre-trained models with high accuracy that I could use for hand detection. My goal was to be able to use this model in projects specific to myself or the organisation I work for, whether it was any object or object, and I achieved this.

I used Kaggle and Google's Open Images Dataset V7 as a dataset. I used the Rsnet34 artificial neural network model for image classification and the YOLOv8 deep learning model for real-time object detection. I developed my project using PyTorch ,Ultralytics (YOLOv8) and OpenCv libraries.

To better explain the working logic of the model, I created a flow diagram that visually explains the data flow and operations of the model.

The aim of this project was to apply my theoretical knowledge on artificial intelligence and train my own model. Moreover, the data sets and models used in the project can be used by other researchers working for similar purposes.

I have a small sample GIF image to show the model I trained using YOLOv8.