Image Processing Methods

Digital Image Processing (DIP) is at the heart of countless innovations and applications that shape our daily lives.
From enhancing the clarity of medical images to enabling advanced features in cameras and smartphones, the algorithms behind these processes play a critical role.
By mastering Digital Image Processing, you gain the ability to transform raw visual data into meaningful and actionable information.

Image Processing is the basis for most vision related systems across industries: Medical, Defense, Production, Sports, etc.
Classic Image Processing is still the way to go in many low level algorithms or implementations on edge devices.

This course teaches advanced and effective methods in Image Processing.

• Image Processing Methods (PDF)

Overview

Digital Image Processing Algorithms: Unveiling the Power of Pixels.
This course dives into the essential techniques and algorithms that form the foundation of modern image processing.
It focuses on learning methods and algorithms both by intuition and by implementation.
The course is based on Hands On approach, learning by exercising on Jupyter Notebooks.

The course:
${\color{lime}\surd}$ Covers the main methods of image processing: Point Wise Operations, Spatial Operations, Image Enhancement, Image analysis, Image Restoration and Geometric Image Processing.
${\color{lime}\surd}$ In depth view of the algorithms accompanied by a detailed code implementation.
${\color{lime}\surd}$ Introduces Edge Preserving Smoothing in depth. Including methods to apply them on large images.
${\color{lime}\surd}$ The course is accompanied by more than 20 notebooks given in a dedicated GitHub repository.

Main Topics

Structure of Digital Image	Structure, Imaging Chain, Noise Model
Point Operations	Pixel Mapping, Histogram, LUT
Spatial and Frequency Domain	Linear & Non Linear Filters, Convolution, Resampling
Optimization, Estimation, Probabilistic Models	Advanced modeling of Image Processing
Image Smoothing	Blurring, Edge Preserving Filters, Variational Models
Image Restoration & Enhancement	Sharpening, Denoising, Deblurring
Image Analysis	Edge Detection, Shape Detection, Multi Scale Analysis
Image Segmentation	Thresholding, Masking, Super Pixels, Image Matting
Image Restoration	Model, Optimization Essentials, Denoising, Deblurring
Feature Based Processing	Feature Extraction, Object Tracking, Patch / Image Registration

Goals

• The participants will be able to describe the structure of a digital image.
• The participants will be able to describe the pipeline of generating a digital image.
• The participants will be able to implement point operations.
• The participants will be able to implement spatial operations.
• The participants will be able to implement image enhancement, image restoration and image analysis algorithms.

Pre Built Syllabus

Day	Subject	Details
1	Course Overview	Motivation, Agenda, Notations
	Scientific Python	NumPy, SciPy, Numba
	Notebook 001	NumPy Arrays
	Notebook 002	Numba and JIT Acceleration
	Notebook 003	SciPy
	Digital Image Structure	Function Model, Color, Imaging Pipeline
	Python Image Processing Eco System	OpenCV, Pillow, SciKit Image
	Point Wise Operations	Mapping, Histogram (Equalization, Transform), Color Spaces, Color to Gray
	Notebook 004	Color Space Transformations
	Notebook 005	Color to Gray by Optimization
2	Spatial Operations	Linear, Non Linear, Shift Invariant, Blurring, Edge Preserving Blur, Interpolation
	Notebook 006	Gaussian Blur by Box Blur
	Notebook 007	Edge Preserving Smoothing: Bilateral Filter, Guided Filter
	Notebook 008	Edge Preserving Smoothing: Weighted Least Squares Filter, Piece Wise Constant Model
	Frequency Domain	The DFT, 2D DFT, Periodic Noise, Resampling in Frequency Domain, The DCT Transform (KLT Approximation)
	Notebook 009	The DFT
	Notebook 010	Efficient Resampling by the DFT
3	Quantization	Sample Quantization, Dithering, Vector Quantization (Lloyd Algorithm, K-Means)
	Notebook 011	Color Quantization by K-Means
	Image Segmentation	Thresholding (Global, Local), Masking, Texture Segmentation (Variance, Entropy), Super Pixels, Variational Methods
	Notebook 012	Thresholding: Local and Global
	Notebook 013	Super Pixels
	Image Matting	Concept, Graphs in Image Processing, Probablistic Model, Matting by Optimization
	Notebook 014	Image Matting
4	Image Analysis	Edge Detection, Shape Detection, Multi Scale Image Analysis, Content Aware Image Resizing
	Notebook 015	Edge and Shape Detection
	Notebook 016	Seam Caring for Content Aware Resizing by Dynamic Programming
	Image Enhancement	Contrast Enhancement, Details Enhancement, Color Enhancement, Color Transfer
	Notebook 017	Local Contrast and Details Enhancement
	Notebook 018	Multi Scale Image Sharpener
	Notebook 019	Recoloring of Image
5	Image Restoration	Intro to Estimation (ML, MAP), Intro to Optimization, Denoising, Deblurring
	Notebook 020	Non Local Means Denoiser
	Notebook 021	Total Variation based Deblurring
	Feature based Image Processing	Feature Detectors, Feature Descriptors, Object Tracking, Image / Patch Registration
	Notebook 022	Corner Detector
	Notebook 023	Multi Scale Optical Flow for Object Detection
	Notebook 024	Football (Soccer) Fields Image Registration

• Image Processing Methods (PDF)

Audience

Algorithm / Software Engineers in the Image Processing, Computer Vision, Deep Learning fields.

Prerequisites

• Mathematical: Linear Algebra (Basic), Probability / Statistics (Basic).
• Programming: Python / MATLAB / Julia.