Browsing by Author "A, Geetha"

Precise identification of liver region from abdominal Computed Tomography-Angiography (CTA) plays an important role in the evaluation of donor for liver transplantation surgery. Nevertheless, the issues like intensity similarity of liver with neighbouring tissues and inter-intra patient liver shape variability; left the task of liver segmentation challenging. Here, we focus on improving the accuracy and reliability of liver donor evaluation system by customising its crucial step - liver segmentation and volume measurement. For achieving this, a Bayesian classifier is iteratively trained with salient features of liver, namely Haralick texture features and spatial information computed from the individual patient dataset. The proposed method is a combination of two techniques namely, advanced region growing and Bayesian classification. The agreement between the proposed method with the manual segmentation was satisfactory with Relative Volume Difference (RVD), Dice Similarity Coefficient (DSC), False-Positive Ratio (FPR), False-Negative Ratio (FNR) with values 8.98, 94.8 ± 1.5, 3.1 ± 2.8 and 5.67 ± 1.8, respectively. Copyright © 2015 Inderscience Enterprises Ltd.

Digital imaging is a powerful technology used to identify illnesses and track the effectiveness of treatment. Although the use of clinical imaging is increasing exponentially, the number of specialists devoted to reviewing it is not growing as quickly. One of the leading causes of cancer-related mortality is the brain tumor, an uncontrolled growth of malignant cells in the brain region. One of the most active research issues in the field is precise brain tumor categorization, which is essential for extracting the correct medical information from magnetic resonance imaging (MRI) scans. A Wiener filter is used in the Extreme learning machine with probabilistic scaling(ELMPS)-based classification model to remove unwanted pixels from the input scan. The photos are classified using an enhanced ELMPS algorithm at the end. Two preprocessing methods, contrast enhancement and skull stripping, are used in the DBN-GWO–based classification model to improve image quality. The fuzzy means clustering (FCM) algorithm is used for picture segmentation. Finally, brain tumors are found in the input photos using an optimized deep belief network (DBN). The grey wolf optimization (GWO) technique enhances the traditional DBN classification performance. © 2024 selection and editorial matter, P. Karthikeyan, Polinpapilinho F. Katina and R. Rajagopal; individual chapters, the contributors.

Augmented Reality (AR) technology has evolved to revolutionize the way consumers interact products, particularly in the fashion and accessory industry. This Project proposes the development of an AR application developed specifically for watches and wrist accessories try-on and fitting. Drawing upon the advancements in marker-based AR technologies, our proposed application aims to provide users with a seamless and immersive experience in virtually trying on various watches and wrist accessories. Utilizing techniques such as marker-based AR integration using, Vuforia, Unity 3D, and Android Studio the application enables users to superimpose virtual watches and accessories onto their wrists in real-time. By leveraging the capabilities of AR-Core SDK and Unity3D, the application facilitates a marker approach hybrid, allowing users to visualize accessories with precision and accuracy. The proposed AR application not only enhances the online shopping experience but also addresses the challenges of traditional in-store try-on methods. Through the convenience of mobile devices, users can explore a diverse range of watches and wrist accessories from renowned brands such as Casio, Sekai, Titan, Sonata, and more. The application may avoid the need for physical visiting the shops. It will also allow for a more user-immersive experience than the online available 2D or 3D images which are usually not to scale or proportional.

The development of aberrant cells, some of which may develop into cancer, results in a brain tumour. Magnetic resonance imaging (MRI) scans frequently reveal brain malignancies. MRI scans are used to identify the abnormal tissue growth in the brain. In several research publications, algorithms for machine learning and deep learning are used to detect brain tumours. It can be used to identify brain tumours quickly and accurately in MRI scans, which makes it simpler to treat patients. These forecasts also help the radiologist act quickly. In the suggested work, preprocessing, segmentation, feature extraction, and classification are all included. An MRI brain image’s undesirable pixels are removed using the Wiener filter during the preprocessing stage. To divide up the data, we applied the fuzzy means clustering (FCM) algorithm. In the second stage, the characteristics of the MRI’s GLCM are extracting the features from the image associated with the MRI brain image. An enhanced extreme learning machine probabilistic scaling is applied in the classification step to categorize the prevailing output image and the interrogation image. The results demonstrate how effective and reliable the suggested methodology is when compared to other recent studies. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

Accurate segmentation of patient’s liver from his/her computed tomography–angiography (CTA) images is the preliminary component for a reliable computerized liver evaluation system. Flawlessness in liver diagnosis relies upon the precision in the segmentation of liver region from all the slices/images in a given patient dataset. Nevertheless, with the challenges like intensity similarity, partial volume effect of liver with its adjacent abdominal organs and liver shape variability across patients, achieving automated optimal liver region segmentation from acquired CT scans is difficult. This paper proposes a semisupervised liver segmentation technique, which adjusts the segmentation parameters for each patient through continuous learning of patient’s CTA dataset properties in a Bayesian level set framework to address all the aforementioned challenges. In this framework, Bayesian probability model with spatial prior is utilized to initiate the level set and to derive an enhanced variable force and edge indication function that helps level set evolution to reach genuine liver boundaries in reduced time. The proposed model has been validated on standard MICCAI liver dataset, producing accuracy score of 79.

In order to determine the possibility that a borrower will not fulfil their financial commitments, financial institutions must evaluate credit risk. Making educated judgments about loan approvals and credit extensions is made easier with the help of the credit risk assessment. Decision Support Systems (DSS) are computer-based solutions that help businesses with difficult decision-making processes by supplying pertinent data and analytical models. DSS is essential for improving the precision and effectiveness of decision-making processes in the context of credit risk evaluation. This paper investigates the use of DSS in credit risk assessment and how it affects the decision-making process. in particular, it looks at data integration, risk assessment models, credit scoring, decision assistance, and portfolio management as important elements of a Credit Risk Evaluation DSS. A complete picture of the borrower's financial situation is possible thanks to the integration of several data sources, including financial statements and credit reports. in order to evaluate this data and find patterns or signs of credit risk, DSS uses sophisticated statistical and machine learning algorithms, which results in more precise risk assessments. Through the use of credit scoring, lenders may divide applicants into several risk groups and make well-informed judgments on loan approvals and interest rates. © 2024 IEEE.

The convergence of cyberspace and space presents new risks and vulnerabilities as technological development ushers in a period in which humankind is more dependent on infrastructure in space. In light of space activities and the necessity of safeguarding space-based systems, this book chapter investigates the application of the Cybersecurity Kill Chain framework. The Cybersecurity Kill Chain, which was originally created to counteract cyber threats on Earth, has been modified to take into account the particulars of environments in orbit. © 2024 River Publishers. All rights reserved.

Edge detection can benefit many different industries and domains, including computer vision, machine learning, image analysis, remote sensing, thermal imaging, pattern recognition, and medical imaging. The technique of determining the borders between several objects or regions in an image is known as edge detection. The edges of an object in a picture serve as the object's limits and can reveal crucial details about the object's size, shape, and position. Since low-resolution images have low pixel densities or pixel values, which muddy the images, detecting edges in them is demanding work. This paper proposes a novel edge-detection approach called EGF (Extended Gaussian Filter) for low-resolution images. EGF utilizes the basic concept of Gaussian filter to find the edges of images. The objective function of EGF is developed to reduce the noise and pixel differentiation in images. The outcomes show that the suggested strategy outperforms the conventional edge detection technique. © 2023 IEEE.

There has been a lot of recent technology advancement in the medical profession. However, in this regard, several time-tested conventional approaches remain popular and effective. X-rays are one technique used to detect bone fractures. A crack doesn't have to be very large for it to go undetected, though. Consequently, they require systems that are both ingenious and well-implemented. The efficiency of the training model, edge detection, segmentation, feature extraction, and preprocessing are all utilized in the suggested method. noise cancellation methods are employed during the preliminary processing of the photos. Edge detection is a method used to identify the boundaries between two things. In order to get more out of an image, segmentation is used. Sharpening an image is part of the feature extraction process that enhances the difference between light and dark areas. Performance is measured using the GRU-SVM Model. The proposed technique is evaluated in comparison to the GRU and SVM models. When compared to older techniques, this one is a huge improvement. © 2023 IEEE.

Internet of Things (IoT) connects various devices of networks, enhances services used to protect against attacks and provide privacy of the user related to all the types of security. This paper analyzes the methods and techniques used in IoT systems with artificial intelligence approach to enhance security. Applying AI algorithms to IoT security allows us to develop smart systems which can detect and block security attacks in real time. Due to the lack of powerful and unified security standards in IoT, an increasing number of IoT devices is vulnerable to threats from malicious attackers and bots. In order to detect attacks and identify abnormal behaviors of smart devices and networks, ML techniques can be used to overcome the issues and challenges. The IoT environment gathers data and analyzes it, and can be done effectively using machine learning, which has the ability to access data, analyze data, and perform decision-making based on data received from IoT devices. This paper addresses the issues which need to be investigated and addressed while implementing the machine learning schemes of security in IoT systems. Respectability, confirmation, and privacy are major principles to be considered to ensure the correspondence between IoT devices. AI offers us a new to solve traditional problems and help us reveal new insights on the field of IoT. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

The technology of augmented reality (AR) has the potential to boost overall productivity in the agricultural sector. It can be used with other technologies including robotic machinery, artificial intelligence (AI), the Internet of Things (IoT), predictive analytics, and precision algorithms. Along with the growing food demand, several microeconomic factors are anticipated to encourage the spread of AR in the agricultural sector. New farmers can now receive creative and safer teaching using augmented reality technology. Young farmers can learn about potential mishaps and mitigate future risks as a result of being electronically exposed to agricultural machinery. Given the facts, it is plausible to assume that augmented reality (AR) technologies are still in use. It has been reasonable to assume that AR technology is still being developed and is not yet sufficiently advanced for agricultural applications. However, there is a significant chance that it will be a big success. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

As the next generation of wireless networks, 6G aims to provide unprecedented levels of performance and connectivity, enabling transformative applications and services. Network virtualization, a concept that has gained prominence in previous network generations, continues to play a pivotal role in the development of 6G networks. In 6G, network virtualization expands beyond traditional network functions virtualization and software-defined networking to encompass novel techniques and paradigms. These include advanced network slicing, edge computing, artificial intelligence–driven orchestration, and dynamic resource management. The integration of edge computing in 6G network virtualization enables the deployment of computation and storage resources closer to end users, reducing latency and improving real-time application performance. In this article, we further explore the concept of virtualization for future generations of wireless networks. © 2024 selection and editorial matter, Dr. Abraham George and G. Ramana Murthy; individual chapters, the contributors.

Medical Decision Support System (MDSS) maps patient information to effective diagnostic and therapeutic pathways. In order to give a robust response to the medical information issue in the situation of missing information, this research presents a comparative examination of various machine learning classifiers for a medical decision-support system. We offer a comparative analysis of an integrated medical decision support system in this paper to help with clinical decisions including the prescription of medications. This study also examines the implementation outcomes brought about by using comparison representations and machine learning techniques to fill in the gaps in the data. © 2023 IEEE.

In the contemporary world, concentrating on predicting the crime analysis is important with increasing criminal & unethical activities all over the world, we are in the need to predict these actions and to formulate countermeasures using statistical/ML techniques. Predictive Policing involves data processing on a large scale to predict and prevent possible future crimes/disturbing activities. This includes utilizing past recorded criminal activities, demographic features, and so on. The method should be able to identify the required patterns accurately and efficiently, and it must use methods that will mesh with law enforcement to make it easier for them to effectively handle these incidents. Using a Recursive Neural Network in Deep Learning (RNNDL) compute confidence rate by giving past predictions less weightage and recent predictions higher weightage, we would be able to figure out crime hotspots with a verified historical dataset of crime records efficiently. ©2024 IEEE.

The segmentation of organs in CT volumes is a prerequisite for diagnosis and treatment planning. In this paper, we focus on liver segmentation from contrast-enhanced abdominal CT volumes, a challenging task due to intensity overlapping, blurred edges, large variability in liver shape, and complex background with cluttered features. The algorithm integrates multidiscriminative cues (i.e., prior domain information, intensity model, and regional characteristics of liver in a graph-cut image segmentation framework). The paper proposes a swarm intelligence inspired edge-adaptive weight function for regulating the energy minimization of the traditional graph-cut model. The model is validated both qualitatively (by clinicians and radiologists) and quantitatively on publically available computed tomography (CT) datasets (MICCAI 2007 liver segmentation challenge, 3D-IRCAD). Quantitative evaluation of segmentation results is performed using liver volume calculations and a mean score of 80.8% and 82.5% on MICCAI and IRCAD dataset, respectively, is obtained. The experimental result illustrates the efficiency and effectiveness of the proposed method. © 2015 Maya Eapen et al.

One of the most interesting and challenging research focuses on pattern recognition and image processing has emerged in recent days is writing in the air. In many different applications, it can improve the interface between a machine and a human and offers a substantial contribution to the development of automated operations. In the field of computer vision, object tracking is considered a key challenge. The method of analyzing a video usually consists of three primary steps: recognizing the object, tracking its movement from frame to frame, and finally evaluating its behavior. Choosing an adequate object representation, selecting tracking features, identifying objects, and tracking them are the four problems taken into consideration for object tracking. Object tracking algorithms are widely used in many real-world applications, including autonomous surveillance, video indexing, and vehicle navigation. This work exploits this gap by developing a motion-to-text converter that may be used as software for wearable intelligent devices that allow writing in the air. The proposed work acts as a recorder of rare gestures. Computer vision will be utilized to track the finger's path. With the generated text, messages, emails, and other kinds of correspondence can all be sent. It will enable effective communication for the deaf. Keywords - object, emoji's, image color, camera © 2023 IEEE.