E-Book, Englisch, 220 Seiten
Kumar / Gupta / Singla Advanced Imaging Applications for Interdisciplinary Engineering
1. Auflage 2026
ISBN: 979-8-89881-456-4
Verlag: De Gruyter
Format: EPUB
Kopierschutz: 0 - No protection
E-Book, Englisch, 220 Seiten
ISBN: 979-8-89881-456-4
Verlag: De Gruyter
Format: EPUB
Kopierschutz: 0 - No protection
Advanced Imaging Applications for Interdisciplinary Engineering is a multidisciplinary exploration of the applications of A.I based imaging technologies and predictive modelling across nanotechnology, quantum imaging, and environmental science.
It covers AI-based image analysis and machine learning applications and innovations in nanoparticle-enhanced diagnostic imaging across modalities such as MRI, CT, and ultrasound. The book also explores quantum imaging techniques, including entanglement-based and ghost imaging, alongside applied computational models for tasks like air quality forecasting. Additional contributions include AI-driven traffic surveillance, image enhancement methods, and blockchain-based healthcare systems for secure data management.
The final sections address environmental studies, including waste analysis and groundwater contamination assessment. Overall, the volume bridges theory and real-world applications across healthcare, environmental monitoring, and intelligent systems.
Key Features:
-Multidisciplinary coverage spanning artificial intelligence, nanotechnology, quantum imaging, and environmental science
-Detailed insights into advanced imaging techniques, including CT, MRI, PET/SPECT, and quantum imaging methods
-Integrated perspectives on machine learning models for predictive analytics, image enhancement, and environmental forecasting
-Explores blockchain-based healthcare frameworks for secure and interoperable data management
-Combines theoretical foundations with practical case studies and real-world applications to provide experimental analysis across medical diagnostics, environmental monitoring, and infrastructure surveillance
Autoren/Hrsg.
Weitere Infos & Material
Investigation of the Imaging Algorithms in Artificial Intelligence
Kanwarpreet Kaur1, *, Payal Patial5, Shonak Bansal3, Meet Kumari2, Muhammed Ali S.A.4
Abstract
Artificial Intelligence (AI) is responsible for the transformation of image processing through cutting-edge imaging algorithms. This chapter delves deeper into AI-based imaging algorithms, such as image classification and pattern recognition. It provides a detailed review of the utilization of machine learning and deep learning models in imaging algorithms. Further, the applications of these AI-driven algorithms are also explored, thus emphasizing the advantages of these models in the real world. Key challenges and opportunities in AI-driven imaging are discussed, offering insights into emerging research directions.
* Corresponding author Kanwarpreet Kaur: Department of Electronics and Communication Engineering, Thapar Institute of Engineering and Technology, Patiala, Punjab, India; E-mail: kpreet2392@gmail.com
INTRODUCTION
Imaging algorithms have always been indispensable for acquiring as well as exchanging information. These algorithms were earlier used for acquiring, transforming, restoring, enhancing, segmenting, and extracting edges of images, but now, with the inception of Artificial Intelligence (AI), there is a significant advancement in the technologies of image algorithms, such as image analysis, image classification, image generation, pattern recognition, and object detection. Despite being the older concept, AI was formalized only in the mid-twentieth
century. The term “Artificial Intelligence” was coined by John McCarthy in 1956. Imaging algorithms in AI involve computer methods, such as processing, analyzing, and interpreting visual information. In the present day, AI-driven imaging algorithms have led to several innovations from healthcare to entertainment. These breakthroughs have become possible only because of the availability of data and computational resources [1-3].
The vital components of AI that are involved in these imaging algorithms are the Machine Learning (ML) and Deep Learning (DL) methods. These algorithms are basically designed to learn the features present in images for performing the tasks of classification, generation, recognition, and detection. In the AI-driven imaging algorithms, the ML and DL models, especially Convolutional Neural Networks (CNNs), are widely employed in various applications. These learning methods are capable of identifying patterns and features for several applications in the image processing domain, which primarily involve image segmentation, image classification and recognition, image enhancement, and image generation [2, 4].
Image segmentation is basically the division of images into meaningful regions, thus making it quite easy to analyze the particular image regions. ML and DL models, such as CNN, AlexNet, and GoogleNet, have gained prominence in performing segmentation, which is important for diagnosis in medical imaging [5, 6]. These algorithms are capable of getting a detailed analysis by segmenting the regions of interest (ROI), thus enhancing the performance metrics. The classification of the images or recognizing the patterns is another field in which the ML and DL models are capable of performing the classification into different categories for the datasets. It can be employed in the majority of imaging applications, such as face recognition, forgery detection, disease detection, etc., which are used in real-time scenarios [7-9].
Further, DL-based superresolution algorithms, such as CNN and residual neural networks, are used for performing the enhancement of images. It involves the enhancement of image resolution, thus making the details prominent. It is beneficial in the case of medical and remote sensing images, in which details are required for performing the analysis to make significant decisions [10, 11]. Further, Generative Adversarial Networks (GANs) are utilized for generating as well as enhancing images, such as the generation of defective images from non-defective images, and performing image-to-image translation for medical images [12-14].
Thus, ML and DL are extensively employed for imaging algorithms in several sectors to increase their efficacy. The capability of the imaging algorithms increases to interpret the data with higher accuracy, thus paving the way for advanced applications. This chapter delves deeper into the historical evolution of AI in imaging algorithms before discussing the ML and DL approaches in imaging algorithms. Afterward, the AI-driven imaging algorithms are explored in various sectors in the subsequent sections before concluding.
Historical Perspective
In the past, the imaging algorithms were focused on simpler applications, such as edge detection and enhancement, based on mathematical approaches, such as transforms. The realization of imaging algorithms using AI dates back to the mid-twentieth century during the exploration of computer vision. Artificial Neural Networks (ANN) and ML models came into the picture at that time to find the optimal solutions for the problems [15, 16].
Towards the end of the twentieth century, learning-based approaches, such as Support Vector Machines (SVM), were introduced. These methods were based on the features considered by the individuals, thus limiting the flexibility. Further, a significant breakthrough was made with the development of the CNN model, which led to the automatic analysis and accurate detection of patterns. A CNN-based method was proposed for recognizing the handwritten characters, which was a solution to the real-world problem. Further, they proposed the Modified National Institute of Standards and Technology (MNIST) dataset for handwritten character recognition [16, 17].
With the advent of DL in the past few decades, particularly with the inception of AlexNet, VGG, ResNets, GoogleNet, etc., there has been rapid advancement in the AI-based applications of image processing in several fields. It has further led to devising solutions for the problems of face recognition, object detection, and disease detection in real-time. In the last decade, the advent of Generative Adversarial Networks (GANs) has further increased the possibilities of more advanced methods not only for performing image analysis, synthesis, and enhancement but also for creating realistic images from scratch [16, 18].
Thus, the evolution of AI-driven imaging has led to a shift from rule-driven approaches to data-driven approaches based on learning. These significant developments have led to the expansion in the implementation of AI across multiple scenarios, ranging from healthcare to industry.
BUILDING BLOCKS OF AI IN IMAGING ALGORITHMS
In broader terms, AI usually refers to the approach that mimics the intelligence of humans. Traditionally, AI was based on two directions, namely, connectionism and computationalism. The former followed a bottom-up approach, which involved the biological neuron-based models. It was based on the emergence of intelligence from learning by experience. In contrast, the latter one does not involve any biological implementation. Instead, it is based on formal logic and reasoning. The former direction involves learning-based AI approaches that further comprise ML and DL. These data-driven approaches are broadly categorized into three learning styles: that is, the basic learning framework, the hybrid learning framework, and learning strategies, as illustrated in Fig. (1). If the model is trained using labeled data to perform the prediction, then it is supervised learning, which is further divided into classification and regression. Here, classification is the categorization of data into different categories, while regression refers to the prediction of continuous values. If the model is trained with...
