Object Detection using YOLOv8, YOLOv5, and Faster R-CNN

This repository contains scripts for performing object detection using the YOLOv5, YOLOv8, and Faster R-CNN models. The scripts are capable of detecting objects in real-time from live video streams captured by a webcam or an external camera.

YOLOv8 Object Detection

The Webcam Object Detection YOLOv8 (cpu).py script utilizes the YOLOv8 model for object detection. It connects to a webcam or an external camera and performs real-time object recognition. The script draws bounding boxes around detected objects along with labels indicating the type of object and its confidence score.

Requirements

• Python 3.x
• OpenCV (cv2)
• NumPy (numpy)
• PIL (Pillow)
• Ultralytics YOLO (ultralytics)

Installation

1. Install required Python packages:

pip install opencv-python numpy Pillow ultralytics

2. Install the YOLO v8 model:

from ultralytics import YOLO
model = YOLO('yolov8n.pt')

3. Usage

python Webcam Object Detection YOLOv8 (cpu).py

By default, the script will use the built-in webcam (camera index 0) for object detection. Press 'q' to exit the program.

Parameters

• camera (int, optional): The camera index. Use 0 for the built-in webcam, or specify 1-n for other connected cameras. Default is 0.
• threshold (float, optional): Confidence threshold for object detection (between 0 and 1). Default is 0.5.
• font_size (int, optional): Font size for object labels. Default is 18.

YOLO v5 Object Detection

The Webcam Object Detection YOLOv5 (cpu).py script utilizes the YOLOv5 model for object detection. It connects to a webcam or an external camera and performs real-time object recognition. The script draws bounding boxes around detected objects along with labels indicating the type of object and its confidence score.

Requirements

• Python 3.x
• OpenCV (cv2)
• NumPy (numpy)
• Torch (torch)
• Ultralytics YOLO (ultralytics)

Installation

1. Install required Python packages:

pip install opencv-python numpy torch torchvision

2. Install the YOLOv5 model:

import torch
model = torch.hub.load('ultralytics/yolov5', 'yolov5s', pretrained=True)

3. Usage

python Webcam Object Detection YOLOv5 (cpu).py

By default, the script will use the built-in webcam (camera index 0) for object detection. Press 'q' to exit the program.

Parameters

• camera (int, optional): The camera index. Use 0 for the built-in webcam, or specify 1-n for other connected cameras. Default is 0.
• threshold (float, optional): Confidence threshold for object detection (between 0 and 1). Default is 0.5.
• font_size (int, optional): Font size for object labels. Default is 18.

Faster R-CNN Object Detection

The PyTorch Webcam Object Detection.py script utilizes the Faster R-CNN model for object detection. It connects to a webcam or an external camera and performs real-time object recognition. The script draws bounding boxes around detected objects along with labels indicating the type of object and its confidence score.

Requirements

• Python 3.x
• OpenCV (cv2)
• NumPy (numpy)
• TorchVision (torchvision)
• PyCOCO Tools (pycocotools)

1. Install required Python packages:

pip install opencv-python numpy torchvision pycocotools

By default, the script will use the built-in webcam (camera index 0) for object detection. Press 'q' to exit the program.

Acknowledgments

This project utilizes the Ultralytics YOLO library for YOLOv8 object detection and TorchVision for Faster R-CNN object detection. YOLO (You Only Look Once) and Faster R-CNN are state-of-the-art, real-time object detection systems that identify multiple objects in images or video streams.

License

This project is licensed under the MIT License - see the LICENSE.md file for details.