Object Detection and Recognition System 🎯

A real-time object detection and recognition system built with Python and YOLOv3, capable of detecting 80+ object categories in both static images and live webcam streams.

---

📌 Project Overview

This project implements two modes of object detection:

• Image Detection — Detects and labels objects in a static image and saves the result
• Real-time Detection — Uses your webcam to detect objects live and saves the output as a video

The system uses the YOLOv3 (You Only Look Once) deep learning model with the COCO dataset (80 object classes), powered by OpenCV's DNN module.

---

🗂️ Project Structure

Object-Detection-and-Recognition-System/
│
├── src/
│   ├── image_detection.py          # Detect objects in a static image
│   └── realtime_detection.py       # Real-time detection via webcam
│
├── model/
│   ├── yolov3.cfg                  # YOLOv3 network configuration
│   ├── coco.names                  # 80 COCO class labels
│   └── yolov3.weights              # ⚠️ Not included — download separately (see below)
│
├── samples/
│   ├── input/                      # Place your sample input images here
│   └── output/                     # Output results saved here
│
├── OUTPUT VIDEOS/                  # Auto-created folder for webcam output videos
│
├── requirements.txt
├── .gitignore
└── README.md

---

⚠️ Download YOLOv3 Weights

The yolov3.weights file (~236MB) is not included in this repository due to file size limits.

Download it from the official YOLO website:

🔗 https://pjreddie.com/media/files/yolov3.weights

After downloading, place it inside the model/ folder:

model/
└── yolov3.weights   ✅

---

⚙️ Setup & Installation

1. Clone the repository

git clone https://github.com/<your-org>/Object-Detection-and-Recognition-System.git
cd Object-Detection-and-Recognition-System

2. Install dependencies

pip install -r requirements.txt

3. Download YOLOv3 weights (see above)

---

🚀 How to Run

🖼️ Image Detection

Detects objects in a static image and displays + saves the result.

1. Place your input image in samples/input/
2. Update the image_path in src/image_detection.py to point to your image
3. Run:

python src/image_detection.py

• Output saved to samples/output/result.jpg

---

📷 Real-time Webcam Detection

Detects objects live via your webcam and saves the video.

python src/realtime_detection.py

• Press ESC to stop
• Output videos auto-saved to OUTPUT VIDEOS/ with timestamped filenames

---

🧠 How It Works

1. Load YOLOv3 model using OpenCV's cv2.dnn.readNet()
2. Preprocess the image/frame into a blob (416×416, normalized)
3. Forward pass through YOLOv3 to get detections
4. Apply NMS (Non-Maximum Suppression) to remove duplicate boxes
5. Draw bounding boxes with class labels and confidence scores
6. Save result image or video

---

🛠️ Tech Stack

Tool	Version
Python	3.10
OpenCV (cv2)	4.9.0.80
NumPy	1.26.4
YOLOv3 Model	COCO (80 classes)
Platform	PyCharm

---

📦 Detectable Object Classes (Sample)

Person, Car, Bicycle, Dog, Cat, Chair, Bottle, Laptop, Phone, Bus, Truck, Bird, and 68 more from the COCO dataset.

---

👩‍💻 Author

Tanuja Devi. M

📄 License

This project is for educational purposes. YOLOv3 model by Joseph Redmon — https://pjreddie.com/darknet/yolo/