Object Detection App

A real-time object detection web app built with YOLOv8 and Streamlit that can detect 80 different object classes from uploaded images.

---

Features

• Image Upload — supports JPG, JPEG, and PNG formats
• YOLOv8n Inference — fast, lightweight model auto-downloaded on first run
• Bounding Boxes — annotated output with confidence scores per detection
• Results Table — per-class breakdown of object, average confidence %, and count
• Plotly Bar Chart — interactive chart of detected object counts
• Confidence Threshold Slider — filter detections in real time (0.1 – 0.9)
• NMS Support — configurable IoU threshold to suppress overlapping duplicate boxes
• Dark Modern UI — custom-styled dark theme with a #63d3a1 accent

---

Tech Stack

Layer	Technology
Language	Python 3.11
Object Detection	YOLOv8 (Ultralytics)
Web Framework	Streamlit
Image Processing	OpenCV, Pillow
Deep Learning	PyTorch, TorchVision
Visualisation	Plotly

---

Run Locally

1. Clone the repository

git clone https://github.com/ezzu9/object-detection-app.git
cd object-detection-app

2. Create a virtual environment (Python 3.11)

python3.11 -m venv venv
source venv/bin/activate        # Windows: venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Launch the app

streamlit run app.py

The app opens at http://localhost:8501. YOLOv8n weights (~6 MB) are downloaded automatically on first run.

---

Model

Property	Value
Architecture	YOLOv8n (nano)
Training Dataset	COCO 2017
Object Classes	80 (people, vehicles, animals, everyday objects, …)
Weights	Auto-downloaded via Ultralytics on first launch

---

Author

Ertaza Manzoor
BSc Artificial Intelligence — Anglia Ruskin University, Cambridge