Название: Intelligent Systems and Applications in Computer Vision
Автор: Nitin Mittal, Amit Kant Pandit, Mohamed Abouhawwash
Издательство: CRC Press
Год: 2024
Страниц: 341
Язык: английский
Формат: pdf (true)
Размер: 42.3 MB

The book comprehensively covers a wide range of evolutionary Computer Vision methods and applications, feature selection and extraction for training and classification, and metaheuristic algorithms in image processing. It further discusses optimized image segmentation, its analysis, pattern recognition, and object detection.

The topic of “computer vision” has grown to encompass a wide range of activities, from gathering raw data to extracting patterns from images and interpreting data. The majority of computer vision jobs have to do with feature extraction from input scenes (digital images) in order to get information about events or descriptions. Computer vision combines pattern detection and image processing. Image understanding comes from the computer vision process. The field of computer vision, in contrast to computer graphics, focuses on extracting information from images. Computer technology is essential to the development of computer vision, whether it is for image quality improvement or image recognition. Since the design of the application system determines how well a computer vision system performs, numerous scholars have proposed extensive efforts to broaden and classify computer vision into a variety of fields and applications, including assembly line automation, robotics, remote sensing, computer and human communications, assistive technology for the blind, and other technologies. Deep Learning (DL) is a member of the AI method family. Artificial Neural Networks (ANNs) get their name from the fact that they receive an input, analyze it, and produce a result. Deep Learning is based on ANN. Because of the massive amount of data generated every minute by digital transformation, AI is becoming more and more popular. The majority of organizations and professionals use technology to lessen their reliance on people.

The large family of algorithms known as “Deep Learning” includes supervised and unsupervised feature learning approaches that include neural networks and hierarchical probabilistic models. Due to their greater performance shown over prior state- of- the- art methods in a number of tasks as well as the volume of complex data from multiple sources, deep learning approaches have recently witnessed an increase in interest. Regarding their applicability in visual understanding, we will concentrate on three one of the key aspects of DL model types in this context: Convolutional Neural Networks (CNN), the “Boltzmann family,” which includes Deep Bolzmann Machines, stacked (denoising) autoencoders, and deep belief networks. Robots used in medical applications have been taught to distinguish between scanned and traditional images.

Features:

Discusses Machine Learning-based analytics such as GAN networks, autoencoders, computational imaging, and quantum computing.
Covers Deep Learning algorithms in computer vision.
Showcases novel solutions such as multi-resolution analysis in imaging processing, and metaheuristic algorithms for tackling challenges associated with image processing.
Highlight optimization problems such as image segmentation and minimized feature design vector.
Presents platform and simulation tools for image processing and segmentation.

The book aims to get the readers familiar with the fundamentals of computational intelligence as well as the recent advancements in related technologies like smart applications of digital images, and other enabling technologies from the context of image processing and computer vision. It further covers important topics such as image watermarking, steganography, morphological processing, and optimized image segmentation. It will serve as an ideal reference text for senior undergraduate, graduate students, and academic researchers in fields including electrical engineering, electronics, communications engineering, and computer engineering.