Image Processing. Theory, Practice, and Deep Learning Approaches (en Inglés) - Fato&#351 Tunay Yarman Vural;Hazal Mo&#287ultay &#214zcan;It&#305r &#214nal Ertu&#287rul

Richly supplemented one-semester textbook on intermediate/advanced image processing

Image Processing introduces a novel approach to image processing methods, combining the foundational and deep learning approaches. It integrates neuroscientific findings with mathematical formalism and practical implementation techniques and seamlessly blends insights from neuroscience and mathematical concepts.

The book is enriched with practical Python programs, allowing readers to run and observe the output of many image processing methods, such as sampling, quantization, interpolation, filtering in spatial and transform domains, histogram operations, morphological operations, boundary extraction, object detection, and image segmentation. Readers can adjust these programs and change various parameters to observe the practical implications of the theoretical representations.

The book is organized into four abstraction levels: Fundamentals of image processing, including the human visual system, mathematical tools for image representation and processing, and color perception with its formal representation. Low-level image processing techniques in the spatial and transform domains, including point operations, histogram techniques, convolutional filters, Fourier, cosine, and Hadamard transforms, multiresolution image analysis, and wavelet transforms. Intermediate-level image processing techniques, including image compression, morphological image processing, image segmentation methods, such as k-means, mean-shift, and normalized cuts, as well as image representation through feature extraction (e.g., polygon approximation, Gabor and SIFT features) and whole-image representation using trees and graphs. High-level image processing techniques with deep learning, including Multi-Layer Perceptrons (MLPs), Artificial Neural Networks, Convolutional Neural Networks (CNNs), Autoencoders (AEs), Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), Diffusion Models (DMs), and Vision Transformers (ViTs) with their applications in image denoising, super-resolution, image colorization, image inpainting, image compression and dimensionality reduction, image segmentation, image-to-text generation, text-to-image generation, and object detection.

This book is an excellent resource for a diverse audience of students and professionals across disciplines who work in designing and implementing image processing algorithms to address both theoretical and practical challenges. Pre-requisites include calculus, probability theory, linear algebra, and programming skills.