Explore chapters and articles related to this topic
Thermal Imaging for Inflammatory Arthritis Evaluation
Published in U. Snekhalatha, K. Palani Thanaraj, Kurt Ammer, Artificial Intelligence-Based Infrared Thermal Image Processing and Its Applications, 2023
U. Snekhalatha, K. Palani Thanaraj, Kurt Ammer
Image segmentation plays an important role in image analysis and may be described as process wherein background objects are separated from the desired section of an image (Preetha et al., 2012). Thermal images depend on the detection of radiation in the infra-red region of the spectrum (Nandhakumar and Aggarwal, 1988). According to black body radiation, each object emanates thermal radiation since its temperature is above absolute zero. For temperatures between ~100 and 1000 K, the wavelength of the thermal rays is in the range of the infrared spectrum. The advantages of infrared imaging cameras over traditional digital cameras are that infrared cameras work also in environments with low levels of visible light. These include security, surveillance, and packaging (Wang et al., 2010, Chen et al., 2003). Typically, a thermal camera is used to acquire the thermal image of the desired body part or ROI such as a subject’s foot, hand, or individual digits. Then, the images are captured and stored using the dedicated software after which the images are pre-processed before being subjected to some form of image segmentation. There are several image segmentation methods that can potentially be used for analyzing the thermal images. These include threshold-based segmentation, edge detection method, k-means clustering, and watershed segmentation (Nandagopan and Haripriya, 2016).
Iris Segmentation in the Wild Using Encoder-Decoder-Based Deep Learning Techniques
Published in Gaurav Jaswal, Vivek Kanhangad, Raghavendra Ramachandra, AI and Deep Learning in Biometric Security, 2021
Shreshth Saini, Divij Gupta, Ranjeet Ranjan Jha, Gaurav Jaswal, Aditya Nigam
Artificial neural networks were used way back in the 1940s initially, but it was only until the 1990s when research dwelled deep into this field [29]. With the availability of digital data sets and development towards computational power, deep learning made huge progress. The very first fully Convolutional Neural Network (CNN) was developed by LeCun et al. [61]. Following the work [61], many researchers started putting a huge effort into developing variations of CNN models, which could give higher performance. In the previous decade, deep learning has brought forth a revolution in the field of image analysis, computer vision, etc. Deep learning is being used for tasks like recognition, segmentation, detection, classification, and so on. Some most recognisable architectures which are used as base model for many derivative works are VGG [97], ResNet [38], GoogLeNet [99], MobileNet [42], DenseNet [44] (not exhaustive list).
Different Techniques Used for Image Processing
Published in Ankur Dumka, Alaknanda Ashok, Parag Verma, Poonam Verma, Advanced Digital Image Processing and Its Applications in Big Data, 2020
Ankur Dumka, Alaknanda Ashok, Parag Verma, Poonam Verma
Image analysis can be used in different applications such as optical character recognition, analysis of medical images, industrial robots, cartography, geology, biometry, and military applications. In terms of optical character recognition, image analysis can be used for sorting of mail, reading of labels, product billing, processing of bank-cheque, reading of text, etc. However in medical images, image analysis can be used for detection of tumors, measurement of organs in terms of their shape and size, analysis of chromosomes, counting of blood cells, etc. (Rajkomar et al., 2019) In industrial robots, it can be used for recognizing and interpretation of objects in a scene and motion control by means of visual feedback. In cartography and geology, it can be used for preparation of maps from photos, plotting of weather maps, exploration of oil, etc. It can also be used in military and other applications as recognition of fingerprint and face, detection and identification of targets, guidance of helicopter and aircraft in times of landing, remote guidance of missiles, etc.
A deep separable neural network for human tissue identification in three-dimensional optical coherence tomography images
Published in IISE Transactions on Healthcare Systems Engineering, 2019
Haifeng Wang, Daehan Won, Sang Won Yoon
Tissue detection through medical imaging is the critical problem when utilizing automated and artificial intelligence (AI) techniques to increase efficiency, reduce healthcare cost, and advance real-time surgery image guidance (Mirnezami and Ahmed, 2018). By using different imaging techniques, such as ultrasound, computed tomography (CT), MRI, mammography, and OCT, physicians can visualize those barely visible tissues inside patients’ bodies. Typically, physicians try to use those medical images to diagnose the patient’s disease manually. Sommerey et al. (2015) tested the use of OCT to identify fat, thyroid, parathyroid, and lymph tissues. The method used in their study was manual assessment, which aimed at matching the image findings to the corresponding histological outcomes. With the rapid development of AI techniques, machine learning approaches have attracted much attention from researchers for automatic image analysis.
A three-dimensional approach to the porous surface of screens
Published in The Journal of The Textile Institute, 2019
A. J. Álvarez, R. M. Oliva, A. Jiménez-Vargas, M. Villegas-Vallecillos
Image analysis can be defined as the extraction of meaningful information from images by means of digital processing techniques (Solomon & Breckon, 2011). Image processing has been proved to be an efficient method of analyzing fabric structures (Jeong & Jang, 2005). These techniques are very important to many applications within the textile industry and their use is widespread and has been used extensively to obtain textile data (Álvarez, Oliva, & Valera, 2012; Cardamone, Damert, Phillips, & Marmer, 2002; Gan, Bickerton, & Battley, 2012; Kang, Choi, Kim, & Oh, 2001; Shin, Cho, Seo, & Kim, 2008). Various techniques have been used including optical scanning, optical microscopy, confocal microscopy, optical coherence tomography, and X-ray microtomography (Sherburn, 2007).
A hybrid lung segmentation algorithm based on histogram-based fuzzy C-means clustering
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2018
Emine Doğanay, Sadık Kara, Hatice Kutbay Özçelik, Levent Kart
The division of an image into meaningful structures, image segmentation, is often an essential step in image analysis, object representation, visualisation and many other image processing tasks. Medical image segmentation is used for finding the exact boundaries of the object, and partitioning an image into multiple segments or regions. Each region or segment consists a set of pixels or voxels (Zhou 2015). Since a great variety of segmentation methods have been proposed in the past decades, image segmentation can be presented within categorisation such edge-based segmentation, region-based segmentation, threshold-based segmentation and clustering etc. However, it is not possible to make a certain categorisation, because different segmentation methods may have possibility to share same properties.