Detectra is a professional AI-powered desktop application designed to detect and document the disappearance of objects in CCTV and video footage. Version 3.0.0 introduces a massive performance and feature update, including hybrid tracking, multi-video support, and advanced forensic visualization.
The system allows users to select an object within a video and intelligently track it across frames using high-performance computer vision. If the object disappears from the scene, Detectra automatically captures high-resolution evidence snapshots, extracts OCR timestamps, and visualizes the object's path for forensic analysis.
- Professional Branding: Custom icons, high-DPI scaling support, and a polished dark-themed interface (Catppuccin).
- Splash Screen: Smooth initial loading experience with real-time status updates for models and dependencies.
- Responsive Canvas: Dynamic video scaling with interactive bounding-box and OCR region selection.
- YOLOv8 + CSRT: Combines state-of-the-art AI detection with OpenCV's CSRT tracker for smooth, reliable, and validated monitoring.
- Resume from Disappearance: Re-draw boxes to track objects that reappear, maintaining forensic continuity.
- Multi-Video Queue: Batch process multiple CCTV files with automated advancement.
- Breadcrumb Investigation: Visualizes the object's trajectory with colored markers (Start: Green, Path: Yellow, Last-seen: Red).
- Precision OCR Selection: Manually define OCR regions to accurately extract timestamps from any DVR layout.
- Night Mode (CLAHE): Integrated contrast enhancement for improved detection in low-light or grainy footage.
- Variable Speed (1x to 30x): Physically seeks through frames for ultra-fast processing of long-duration footage without UI lag.
- Standalone EXE: Packaged as a single-file portable Windows executable with localized model storage.
- OS: Windows 10 or 11 (64-bit)
- RAM: 8 GB (16 GB Recommended)
- GPU: NVIDIA GPU with CUDA support recommended (but runs on CPU)
- Dependencies: The standalone version requires an internet connection on the first launch only to download models.
- Download
Detectra_v3.0.0.exefrom the Releases page. - Run the file directly. No installation is required.
- Clone the repo:
git clone https://github.com/NEIL-DANIEL-A/Detectra.git - Create virtual environment:
python -m venv venv - Activate:
venv\Scripts\activate - Install requirements:
pip install -r requirements.txt - Run:
python main.py
| File | Description |
|---|---|
main.py |
Application entry point, UI management, and Splash Screen. |
tracker.py |
Core logic for Hybrid Tracking (YOLOv8+CSRT) and OCR. |
requirements.txt |
Python dependencies for the developer environment. |
icon.ico |
High-resolution application branding. |
- Python 3.10+ & Tkinter — Core application framework
- Ultralytics YOLOv8 — State-of-the-art object detection
- OpenCV & CSRT — Advanced video processing and tracking
- EasyOCR — Optical Character Recognition for forensic timestamps
- PyInstaller — Secure executable distribution
Current version: v3.0.0
- Hybrid Tracker: Replaced basic YOLO tracking with a validated YOLO + CSRT engine.
- Multi-File Queue: Support for batch processing multiple videos.
- Path Visualization: Real-time breadcrumb trails and forensic path capture.
- OCR Region Tool: Custom selection of timestamp areas for better accuracy.
- Resume Feature: Ability to continue tracking from a disappearance point.
- CLAHE Preprocessing: Improved visibility for dark and low-contrast footage.
This project is developed for educational and research purposes.