Skip to content

object_detection_system

A comprehensive computer vision application for detecting, segmenting, and tracking objects in images and videos, with optional LLM integration for enhanced object recognition.

View on GitHub →

Overview

Languages: Python

Documentation

Object Detection and Tracking System

A comprehensive computer vision application for detecting, segmenting, and tracking objects in images and videos, with optional LLM integration for enhanced object recognition.

Features

  • Object Detection: Identify objects in images and videos using state-of-the-art models (YOLOv8, TensorFlow, CoreML)
  • Instance Segmentation: Generate precise object outlines using Segment Anything Model (SAM)
  • Object Tracking: Track objects across video frames with ByteTrack or SORT algorithms
  • LLM Integration: Enhanced object recognition using Ollama (local) or Cloud APIs
  • Multi-Format Support: Process various image and video formats
  • Hardware Optimization: Automatic detection and utilization of available hardware (CUDA, MPS on Apple Silicon)
  • User Authentication: Secure login system with user management
  • Plugin System: Extensible architecture through plugins
  • Rich Visualization: Interactive display with zooming, panning, and object highlighting

System Requirements

  • Python 3.8 or higher
  • PyTorch 2.0 or higher
  • OpenCV 4.7 or higher
  • PyQt5 5.15 or higher
  • 4GB RAM minimum (8GB+ recommended)
  • NVIDIA GPU with CUDA support (optional, for acceleration)
  • Apple Silicon (M1/M2/M3) for MPS acceleration (optional)

Installation

  1. Clone the repository: 

    git clone https://github.com/yourusername/object-detection-system.git
cd object-detection-system

  2. Install dependencies: 

    pip install -r requirements.txt

  3. Download models:

For YOLOv8: 

mkdir -p data/models
# Download YOLOv8n (small)
wget -O data/models/yolov8n.pt https://github.com/ultralytics/assets/releases/download/v0.0.0/yolov8n.pt

For SAM (Segment Anything Model): 

# Download SAM ViT-B
wget -O data/models/sam_vit_b_01ec64.pth https://dl.fbaipublicfiles.com/segment_anything/sam_vit_b_01ec64.pth

Usage

Starting the Applica

[View full README on GitHub]