shaaleissa/traffic_detection

Comparative Analysis of YOLOv8 and YOLOv5 Model Performance in Object Detection Project

Overview

This repository provides a comprehensive comparison between YOLOv8 and YOLOv5 models in the realm of object detection. It includes various Jupyter notebooks that detail the processes of detection, training, evaluation, and data preprocessing.

Repository Structure

Notebooks Directory

This directory is organized into several subdirectories, each containing Jupyter notebooks for specific tasks:

Detection Directory

YOLOv5_detection.ipynb YOLOv8_detection.ipynb These notebooks are dedicated to object detection using YOLOv5 and YOLOv8 models, respectively. Training and Evaluation Directory

YOLOv5_training_evaluation.ipynb YOLOv8_training_evaluation.ipynb These notebooks cover the training and evaluation phases for both YOLOv5 and YOLOv8 models. Data Preprocessing

data_preprocessing.ipynb Contains the exploratory data analysis and preprocessing steps. Predict Directory Contains results from object detection experiments using both YOLOv5 and YOLOv8 models.

Models Directory

Stores the best-performing weights for both YOLOv5 and YOLOv8 models, trained on custom datasets.

Images Directory

Includes all images utilized in the object detection experiments.

Usage Instructions

Training and Evaluation: The training and evaluation processes were conducted using Kaggle notebooks. Ensure you have the necessary environment to run these notebooks. Detection: The detection notebooks can be run locally, provided that the model weights are placed in the models directory.

Getting Started

To get started with this project:

• Clone the repository.
• Ensure you have Jupyter Notebook installed, or use Kaggle notebooks for training and evaluation purposes.
• For detection tasks, download the pre-trained weights and place them in the models directory.
• Run the notebooks to perform detection, training, evaluation, and data preprocessing.