🎯 Python-Based Object Detector using OpenCV

A Python-based object detection application for images and videos using the OpenCV Deep Neural Networks (DNN) module with GoogLeNet and YOLOv3 for real-time, high-accuracy detection.
This project leverages deep learning inference to classify and detect objects with bounding boxes and probability scores, using pre-trained models on the COCO dataset.

---

🧠 Features

• 🖼 Image Object Detection – Detects multiple objects in an image and displays their bounding boxes with confidence scores.
• 🎥 Video Object Detection – Detects objects frame-by-frame with probability scores.
• 🧮 Color Channel Analysis – Displays different color channels from the image.
• ⚡ YOLOv3 Neural Network – Divides images into regions and predicts bounding boxes and class probabilities for each region.
• 📦 Pre-Trained Models – Uses GoogLeNet & YOLOv3 models trained on the COCO dataset.

---

🛠 Tech Stack

Component	Technology
Language	Python
Core Library	OpenCV
Deep Learning	OpenCV DNN module
Models Used	GoogLeNet, YOLOv3
Dataset	COCO dataset

---

📚 Learning Objectives

• Deep learning for OpenCV
• Viewing images and videos in OpenCV
• Working with blobs in the DNN module
• Image classification & video classification
• YOLOv3-based object detection

---

🗂 Folder Structure

/opencv-object-detector
│
├── /models            # Pre-trained YOLOv3 & GoogLeNet models
├── /images            # Sample images for detection
├── /videos            # Sample videos for detection
├── detector.py        # Main Python script for object detection
├── requirements.txt   # Python dependencies
└── README.md

🚀 Getting Started

1️⃣ Clone the Repository

git clone https://github.com/Iamdk25/opencv-object-detector.git
cd opencv-object-detector

2️⃣ Install Dependencies

pip install -r requirements.txt

3️⃣ Download Pre-Trained Models • YOLOv3 Weights • YOLOv3 Config • GoogLeNet pre-trained model from the OpenCV model zoo

Place these files inside the /models directory.

4️⃣ Run Object Detection on an Image

python detector.py --image images/sample.jpg

5️⃣ Run Object Detection on a Video

python detector.py --video videos/sample.mp4

🏆 Completed Functionality

•	✅ Detect objects in images & display color channels
•	✅ Detect objects in videos with probability scores
•	✅ YOLOv3-based custom object detection
•	✅ Bounding box & confidence prediction per region
•	✅ Inference-based classification for images & videos

⸻

📜 License

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

⸻

👨‍💻 Author

Divyarajsinh Karmariya
💼 Computer Science Student @ University of South Florida
🔗 Portfolio • 🧠 GitHub • 💼 LinkedIn