👁️ Computer Vision & Real-Time Processing Suite

📖 Executive Summary

This repository hosts a professional-grade suite of Computer Vision modules designed to solve real-world image processing and analysis challenges. Built with Python and OpenCV, this project demonstrates advanced proficiency in algorithmic image manipulation, real-time video analytics, and deep learning model integration.

Unlike standard scripts, these tools are wrapped in robust Tkinter GUIs, featuring multi-threaded processing for responsiveness, JSON-based persistent configuration, and error-resilient architectures. This collection serves as a powerful reference for implementing production-ready computer vision systems.

---

🛠️ Technical Competency Matrix

This project demonstrates mastery across the following key technical domains:

Domain	Skills Illustrated
Advanced Computer Vision	Feature Matching (SIFT/ORB), Homography, Watershed Segmentation, Hough Transforms, Optical Flow.
Deep Learning Integration	Deployment of ONNX models (YuNet) for high-performance Face Detection within standard OpenCV pipelines.
Software Engineering	Multi-threading (GUI responsiveness), Exception Handling, Modular Design, User Configuration Management (JSON).
GUI Development	Complex Tkinter interfaces with tabbed navigation, real-time canvas rendering, and dynamic parameter control.
Mathematical Foundations	Matrix operations, geometrical transformations, histogram analysis, and statistical image filtering.

---

📂 Module Breakdown

1. 🤖 Deep Learning Face Detection

File: face_Detection_OpenCV.py A production-ready face detection application utilizing the YuNet deep neural network (via ONNX).

• Capabilities: Real-time detection from webcam or video files; static image analysis.
• Key Features:
  • Adjustable Confidence & NMS (Non-Maximum Suppression) thresholds.
  • Visual Debugging: Renders bounding boxes and 5-point facial landmarks.
  • Persistence: Automatically saves/loaded user preferences via face_detection_settings.json.
  • Performance: Multi-threaded video loop ensures the UI never freezes.

2. 🎯 Real-Time Object Tracking

File: Object_trackeng_OpenCV.py A versatile tracking suite implementation supporting the full range of OpenCV tracking APIs.

• Algorithms Supported: KCF, CSRT (State-of-the-art), MIL, BOOSTING, MEDIANFLOW, MOSSE, TLD.
• Intelligence: Automatically handles legacy vs. modern OpenCV version differences.
• Usage: Select an ROI (Region of Interest) with the mouse and track it robustly through occlusions and movement.

3. 🧩 Intelligent Image Segmentation

File: image_Segmentation_OpenCV.py An interactive laboratory for isolating objects from backgrounds.

• Methods:
  • Watershed Algorithm: Marker-based segmentation for touching objects.
  • Adaptive Thresholding: Gaussian-weighted local thresholding for varied lighting.
  • Global Thresholding: Standard binary segmentation.
• Controls: Real-time sliders for kernel sizes, threshold factors, and block sizes.

4. 📐 Geometry & Feature Analysis

• Hough Line Detection (Hough_Transform_for_LineDetection.py): Detects structural lines in images using probabilistic Hough transforms. Essential for lane detection and document scanning.
• Feature Mapping (feature_Mapping_OpenCV.py): Implements SIFT/SURF/ORB descriptors to find homographies and match objects across different perspectives.
• Image Stitching (image_Stitching_OpenCV.py): Automates panorama creation by stitching multiple overlapping photos.

5. 🎨 Image Processing & Artistry

• Pencil Sketching (Pencil_sketch_conversion_OpenCV.py): Converts photorealistic images into artistic pencil sketches using gaussian differences and division blending.
• Morphological Ops (morphological_transformations_OpenCV.py): Interactive tool for Erosion, Dilation, Opening, and Closing—critical for noise removal in binary images.
• Live Grayscale (live_GrayScale_Conversion.py): High-performance video stream color space conversion.

---

⚙️ Installation & Setup

Prerequisites

• Python 3.8+
• Webcam (for live modules)

Dependencies

Install the required packages using pip:

pip install opencv-python opencv-contrib-python numpy pillow scipy

Note: opencv-contrib-python is recommended for full feature access (like certain tracking algorithms).

Model Setup

For the Face Detection module, ensure the YuNet ONNX model is present in the root directory:

• face_detection_yunet_2023mar.onnx

---

🚀 Usage Guide

Running FLASK / GUI Applications

Most modules are standalone scripts. Run them directly via Python:

To launch Face Detection:

python face_Detection_OpenCV.py

Navigate the tabs to switch between "Image" and "Video" modes. Adjust thresholds in "Settings".

To launch Object Tracking:

python Object_trackeng_OpenCV.py

Follow the CLI prompts to select a tracker (e.g., KCF or CSRT), then draw a box around the object you wish to track.

---

📂 Project Structure

.
├── 📄 face_Detection_OpenCV.py          # AI Face Detection App
├── 📄 Object_trackeng_OpenCV.py         # Multi-algo Object Tracker
├── 📄 image_Segmentation_OpenCV.py      # Watershed/Thresholding Tool
├── 📄 Hough_Transform_for_LineDetection.py
├── 📄 feature_Mapping_OpenCV.py
├── 📄 image_Stitching_OpenCV.py
├── 📄 Pencil_sketch_conversion_OpenCV.py
├── ⚙️ face_detection_settings.json      # Config file (Auto-generated)
├── 🧠 face_detection_yunet_2023mar.onnx # Pre-trained Model
└── 📂 sample_files/                     # Test footage (Airport, Crowd, etc.)

---

🤝 Contribution & License

This project is open-source and available for educational and commercial modification.

• License: MIT License
• Contributions: Pull requests designed to add new OpenCV algorithms (e.g., YOLOv8 integration, Optical Character Recognition) are highly encouraged.

---

Developed with precision for Computer Vision Engineering roles.