👁️ Computer Vision: Edge & Shape Detection Pipeline (C++ / Qt)

📌 Project Overview

A comprehensive desktop application developed in C++ using the Qt Framework. This project implements core Computer Vision algorithms from scratch, focusing on edge detection, geometric shape recognition, and contour tracking through mathematical energy minimization.

⚙️ Core Algorithms Implemented

1. Canny Edge Detector

Implemented the complete 5-step Canny edge detection pipeline from scratch:

• Gaussian Blur for noise reduction.
• Sobel Gradients (Magnitude & Angle computation).
• Non-Maximum Suppression (Edge thinning).
• Double Thresholding.
• Edge Tracking by Hysteresis.

2. Hough Transform (Shape Detection)

• Hough Circles: Utilizes a highly optimized accumulator to detect circular shapes.
• Hough Ellipses & Lines: Extracts generalized geometric shapes from binary edge maps by mapping pixels to parameter space.

3. Active Contours (Snakes)

Implemented an iterative energy-minimizing spline model for contour tracking and image segmentation. The contour actively snaps to object boundaries by minimizing the total energy: $$E_{total} = \alpha E_{cont} + \beta E_{curv} + \gamma E_{img}$$ Where $E_{cont}$ enforces continuity, $E_{curv}$ enforces smoothness, and $E_{img}$ pulls the snake toward image edges (computed via Gaussian blurring and Canny edge distance transforms).

---

📸 Application Output Gallery

1. Hough Transform: Multi-Shape Detection (Lines & Circles)

Simultaneous detection and overlay of multiple geometric primitives.

2. Hough Transform: Circle Detection

Robust detection of circular shapes using a custom accumulator.

3. Hough Transform: Ellipse Detection

Extracting elliptical structures by mapping pixels to the required parameter space.

4. Active Contours (Snakes) Tracking

Visualizing the iterative process of the contour snapping to the object's boundaries by minimizing energy functions.

---

🛠️ Tech Stack & Architecture

• Language: C++ (Object-Oriented Design)
• GUI Framework: Qt (Widgets / UI Designer)
• Build System: CMake
• Libraries: OpenCV (Used strictly for matrix operations cv::Mat and basic rendering, while core algorithms are implemented from scratch).

🚀 Build & Run Instructions

Prerequisites

• CMake (Version 3.16 or higher)
• Qt5 or Qt6 libraries
• C++17 Compiler

Steps

1. Clone the repository:

   git clone [https://github.com/YourUsername/CV-Edge-Shape-Detection.git](https://github.com/YourUsername/CV-Edge-Shape-Detection.git)