Object Detection App (YOLOv11)

1 Project Overview

This project demonstrates real-time object detection using Ultralytics YOLOv11.
It includes a desktop GUI (Tkinter), suitable for local GPU/CPU use.

1.1 Key Features

  • Detects objects in images, videos, and live camera feeds. (This report only showcases image results for clarity and simplicity.)
  • Side-by-side Original vs Detected views for easy comparison.
  • Counts selected classes such as person, car, truck, bicycle.

2 Visual Results

2.1 Before and After: Image Detection

The model identifies objects and overlays bounding boxes with class labels and confidence scores.

2.1.1 Example 1

  • Before: Original uploaded image.
  • After: Detected objects highlighted with bounding boxes.


2.1.2 Example 2

  • Before: Original frame from video.
  • After: Detected vehicles and pedestrians annotated.



### Example 3 - Before: Original frame from video.
- After: Detected vehicles and pedestrians annotated.


3 Quick Start (Local)

3.1 Tkinter Desktop App

# 1) Create & activate a virtual environment
python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS/Linux
# source .venv/bin/activate

# 2) Install requirements
pip install -r requirements.txt

# 3) Run desktop GUI
python main.py

4 Technologies Used

  • Python (app logic)
  • Torch & NumPy (deep learning computations)
  • Ultralytics YOLOv11 (object detection model)
  • OpenCV (video/image handling)
  • Pillow (image processing)
  • Tkinter (desktop GUI)

5 Future Improvements

  • Add real-time detection dashboard in Streamlit.
  • Introduce custom class training for specific use cases.
  • Optimize inference speed for low-power devices.
  • Add export to Excel/CSV for detection logs.