IR2126 - Computer Vision

This repository contains all practical assignments developed for IR2126 - Computer Vision, a 3rd year core subject in the Bachelor’s Degree in Robotic Intelligence at Universitat Jaume I.

The course offers theoretical foundations and practical applications in digital image processing and computer vision. Each task is implemented in Python and accompanied by a detailed PDF report outlining the methodology, parameters used, and results.

---

🎯 Learning Goals

• Use spatial and frequency domain filters for image enhancement
• Apply geometric transformations on digital images
• Detect edges, contours, and image features
• Perform object segmentation using morphological techniques
• Understand shape descriptors and introductory 3D vision principles

---

📦 Folder Contents

• 🐍 Python scripts for each practical assignment
• 🖼️ Input and output sample images
• 📑 PDF report with methods, parameter tuning, and result analysis

---

🧪 Practicals Overview

🖍️ Practice 1 – Introduction to Image Processing

• RGB channel extraction and visualization
• Image alignment and composition
• JPEG compression analysis
• Creation of animated GIFs

---

🌗 Practice 2 – Contrast Adjustment & Equalization

• Contrast enhancement using linear scaling and histogram equalization
• Comparison between grayscale and RGB enhancement
• Histogram analysis and kernel size impact

---

📐 Practice 3 – Geometric Transformations

• Rotation and affine transforms using transformation matrices
• Chaining multiple transformations
• Mesh warping using piecewise affine techniques

---

🔬 Practice 4 – Spatial Filtering

• Noise simulation (Gaussian, salt & pepper)
• Comparing 2D convolution with separable 1D filters
• Image sharpening via high-pass filtering

---

🎚️ Practice 5 – Frequency Domain Filtering

• Fourier-based convolution filtering
• Implementation of high-, low-, and band-pass filters
• Spatial vs frequency domain comparison in terms of efficiency

---

🧩 Practice 6 – Edge & Feature Detection

• Edge detection: Sobel, Canny
• Hysteresis thresholding
• Hough transform for line/circle detection
• Corner detection using Harris, Moravec, and Foerstner methods

---

🧬 Practice 7 – Segmentation & Morphology

• Threshold-based binary segmentation
• Morphological operations: erosion, dilation, opening, closing
• Shape classification based on roundness (e.g., coin detection: €1 vs 10c)

---

🧰 Tools & Libraries

• Python 3
• numpy, scikit-image, matplotlib, scipy
• Developed in PyCharm

---

👨‍🏫 Instructor

Pedro García Sevilla

---

📂 Browse each folder to explore the code, visual results, and full documentation for every practical task.