Kannan Ramakrishnan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Kannan Ramakrishnan has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
- Rathimala Kannan
PSG College of Technology
7 shared publications - Mohamed Uvaze Ahamed Ayoobkhan
American University of Technology, Uzbekistan
3 shared publications Rathimala Kannan
Information Technology Department Multimedia University Cyberjaya Malaysia
3 shared publicationsEswaran Chikkannan
Multimedia University
2 shared publicationsHway Boon Ong
Department of Economics, Faculty of Management, Multimedia University, Cyberjaya, Selangor, 63100, Malaysia
2 shared publicationsIvan Zhi Wei Wang
Faculty of Management, Multimedia University, Cyberjaya, Selangor, 63100, Malaysia
2 shared publications
Latest Publications
- 1 May 2026
- Published by Springer Nature
- p. 65-74
AI‑Based Algorithmic Predictions of Purchase Intention, and Loyalty: A Multi‑country Study
- 1 May 2026
- Published by Springer Nature
- p. 65-74
AI‑Based Algorithmic Predictions of Purchase Intention, and Loyalty: A Multi‑country Study
- Open Access
- 27 May 2025
- Published by MDPI AG in Information
- Vol. 16 (6) , 451
Deep Learning Techniques for Lung Cancer Diagnosis with Computed Tomography Imaging: A Systematic Review for Detection, Segmentation, and Classification
Background/Objectives: Lung cancer is a major global health challenge and the leading cause of cancer-related mortality, due to its high morbidity and mortality rates. Early and accurate diagnosis is crucial for improving patient outcomes. Computed tomography (CT) imaging plays a vital role in detection, and deep learning (DL) has emerged as a transformative tool to enhance diagnostic precision and enable early identification. This systematic review examined the advancements, challenges, and clinical implications of DL in lung cancer diagnosis via CT imaging, focusing on model performance, data variability, generalizability, and clinical integration. Methods: Following the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we analyzed 1448 articles published between 2015 and 2024. These articles are sourced from major scientific databases, including the Institute of Electrical and Electronics Engineers (IEEE), Scopus, Springer, PubMed, and Multidisciplinary Digital Publishing Institute (MDPI). After applying stringent inclusion and exclusion criteria, we selected 80 articles for review and analysis. Our analysis evaluated DL methodologies for lung nodule detection, segmentation, and classification, identified methodological limitations, and examined challenges to clinical adoption. Results: Deep learning (DL) models demonstrated high accuracy, achieving nodule detection rates exceeding 95% (with a maximum false-positive rate of 4 per scan) and a classification accuracy of 99% (sensitivity: 98%). However, challenges persist, including dataset scarcity, annotation variability, and population generalizability. Hybrid architectures, such as convolutional neural networks (CNNs) and transformers, show promise in improving nodule localization. Nevertheless, fewer than 15% of the studies validated models using multicenter datasets or diverse demographic data. Conclusions: While DL exhibits significant potential for lung cancer diagnosis, limitations in reproducibility and real-world applicability hinder its clinical translation. Future research should prioritize explainable artificial intelligence (AI) frameworks, multimodal integration, and rigorous external validation across diverse clinical settings and patient populations to bridge the gap between theoretical innovation and practical deployment.
- 27 May 2025
- Published by MDPI AG in Information
- Vol. 16 (6) , 451
Open AccessDeep Learning Techniques for Lung Cancer Diagnosis with Computed Tomography Imaging: A Systematic Review for Detection, Segmentation, and Classification
Background/Objectives: Lung cancer is a major global health challenge and the leading cause of cancer-related mortality, due to its high morbidity and mortality rates. Early and accurate diagnosis is crucial for improving patient outcomes. Computed tomography (CT) imaging plays a vital role in detection, and deep learning (DL) has emerged as a transformative tool to enhance diagnostic precision and enable early identification. This systematic review examined the advancements, challenges, and clinical implications of DL in lung cancer diagnosis via CT imaging, focusing on model performance, data variability, generalizability, and clinical integration. Methods: Following the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we analyzed 1448 articles published between 2015 and 2024. These articles are sourced from major scientific databases, including the Institute of Electrical and Electronics Engineers (IEEE), Scopus, Springer, PubMed, and Multidisciplinary Digital Publishing Institute (MDPI). After applying stringent inclusion and exclusion criteria, we selected 80 articles for review and analysis. Our analysis evaluated DL methodologies for lung nodule detection, segmentation, and classification, identified methodological limitations, and examined challenges to clinical adoption. Results: Deep learning (DL) models demonstrated high accuracy, achieving nodule detection rates exceeding 95% (with a maximum false-positive rate of 4 per scan) and a classification accuracy of 99% (sensitivity: 98%). However, challenges persist, including dataset scarcity, annotation variability, and population generalizability. Hybrid architectures, such as convolutional neural networks (CNNs) and transformers, show promise in improving nodule localization. Nevertheless, fewer than 15% of the studies validated models using multicenter datasets or diverse demographic data. Conclusions: While DL exhibits significant potential for lung cancer diagnosis, limitations in reproducibility and real-world applicability hinder its clinical translation. Future research should prioritize explainable artificial intelligence (AI) frameworks, multimodal integration, and rigorous external validation across diverse clinical settings and patient populations to bridge the gap between theoretical innovation and practical deployment.
- 22 April 2025
- Published by Springer Nature
- p. 261-275
Financial Data-Driven Classification of Manufacturing Companies: A Case Study of the Malaysian Industry
- 22 April 2025
- Published by Springer Nature
- p. 261-275
Financial Data-Driven Classification of Manufacturing Companies: A Case Study of the Malaysian Industry
- Open Access
- 14 February 2025
- Published by Multimedia University in Journal of Informatics and Web Engineering
- Vol. 4 (1) , 168-183
Early Identification of Parkinson's Disease Using Time Frequency Analysis on EEG Signals
Parkinson's Disease (PD) is a progressive neurological disorder. It affects movement and can significantly impact quality of life. Early and accurate diagnosis is crucial for effective management and intervention. Traditional diagnostic methods can be time-consuming and less effective in the early stages of the disease. This study aims to develop an automated approach for identifying PD using time-frequency image analysis of electroencephalogram (EEG) signals. The goal is to enhance diagnostic accuracy and efficiency, facilitating early detection. EEG signals, often contaminated with artifacts such as eye blinks and muscle movements etc., were first cleaned. Time-frequency images were then plotted from the cleaned signals, and Event-Related Spectral Perturbation (ERSP) plots were extracted. A customized deep learning model was employed to classify the ERSP plots, distinguishing PD patients from healthy controls. The deep learning model achieved an accuracy of 94.64% in separating PD patients from healthy controls. The approach demonstrated robustness against common EEG artifacts, ensuring reliable PD detection. The model's architecture was specifically designed to handle the complexities of EEG data, making it a powerful tool for PD classifications. This study highlights the potential of integrating deep learning with EEG analysis to explore PD diagnosis. The proposed method is faster and more accurate than traditional approaches, enabling early detection and timely intervention. By reducing the time required for analysis and enhancing diagnostic accuracy, this approach can significantly improve patient outcomes and support better management of Parkinson's Disease.
- 14 February 2025
- Published by Multimedia University in Journal of Informatics and Web Engineering
- Vol. 4 (1) , 168-183
Open AccessEarly Identification of Parkinson's Disease Using Time Frequency Analysis on EEG Signals
Parkinson's Disease (PD) is a progressive neurological disorder. It affects movement and can significantly impact quality of life. Early and accurate diagnosis is crucial for effective management and intervention. Traditional diagnostic methods can be time-consuming and less effective in the early stages of the disease. This study aims to develop an automated approach for identifying PD using time-frequency image analysis of electroencephalogram (EEG) signals. The goal is to enhance diagnostic accuracy and efficiency, facilitating early detection. EEG signals, often contaminated with artifacts such as eye blinks and muscle movements etc., were first cleaned. Time-frequency images were then plotted from the cleaned signals, and Event-Related Spectral Perturbation (ERSP) plots were extracted. A customized deep learning model was employed to classify the ERSP plots, distinguishing PD patients from healthy controls. The deep learning model achieved an accuracy of 94.64% in separating PD patients from healthy controls. The approach demonstrated robustness against common EEG artifacts, ensuring reliable PD detection. The model's architecture was specifically designed to handle the complexities of EEG data, making it a powerful tool for PD classifications. This study highlights the potential of integrating deep learning with EEG analysis to explore PD diagnosis. The proposed method is faster and more accurate than traditional approaches, enabling early detection and timely intervention. By reducing the time required for analysis and enhancing diagnostic accuracy, this approach can significantly improve patient outcomes and support better management of Parkinson's Disease.
- Open Access
- 20 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 16252-16262
VoxelSky-3D: A Weather Radar Visualization Prototype for Air Traffic Control
This paper presents VoxelSky-3D, a new 3D weather radar visualization prototype for civil aviation air traffic control. While previous research has explored text-based, image-based, and some three-dimensional meteorological displays, none have fully integrated voxelization techniques to deliver real-time, comprehensive 3D weather visualizations specifically tailored for Air Traffic Control Officers (ATCOs). The lack of traditional 2D weather radar displays vertical atmospheric structures for ATCOs, which are critically important for safety and efficiency during flight. VoxelSky-3D is an auxiliary system that uses voxelization to transform weather radar data into 3D voxel-based representations. The methods used in this study include the conversion of coordinates, voxelizing of the geographic coordinate system, and integration with Geographic Information Systems for rendering via the Web Graphics Library. The experimental results verified that VoxelSky-3D is an effective application for visualizing weather radar data in 3D, which directly provides ATCOs with intuitive views of the vertical atmospheric conditions. The performance test results revealed that the system could build real-time complex 3D weather scenes and provide high frame rates under relatively high volumes of incoming data. The VoxelSky-3D prototype is a considerable step forward in the field of meteorological data visualization for air traffic control, which improves the situational awareness and decision-making capabilities of ATCOs.
- 20 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 16252-16262
Open AccessVoxelSky-3D: A Weather Radar Visualization Prototype for Air Traffic Control
This paper presents VoxelSky-3D, a new 3D weather radar visualization prototype for civil aviation air traffic control. While previous research has explored text-based, image-based, and some three-dimensional meteorological displays, none have fully integrated voxelization techniques to deliver real-time, comprehensive 3D weather visualizations specifically tailored for Air Traffic Control Officers (ATCOs). The lack of traditional 2D weather radar displays vertical atmospheric structures for ATCOs, which are critically important for safety and efficiency during flight. VoxelSky-3D is an auxiliary system that uses voxelization to transform weather radar data into 3D voxel-based representations. The methods used in this study include the conversion of coordinates, voxelizing of the geographic coordinate system, and integration with Geographic Information Systems for rendering via the Web Graphics Library. The experimental results verified that VoxelSky-3D is an effective application for visualizing weather radar data in 3D, which directly provides ATCOs with intuitive views of the vertical atmospheric conditions. The performance test results revealed that the system could build real-time complex 3D weather scenes and provide high frame rates under relatively high volumes of incoming data. The VoxelSky-3D prototype is a considerable step forward in the field of meteorological data visualization for air traffic control, which improves the situational awareness and decision-making capabilities of ATCOs.
- Open Access
- 22 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 180792-180814
Optimizing Machine Learning Models for Accessible Early Cognitive Impairment Prediction: A Novel Cost-Effective Model Selection Algorithm
Cognitive impairment and dementia-related diseases develop several years before moderate or severe deterioration in cognitive function occurs. Nevertheless, most dementia cases, especially in low- and middle-income countries, remain undiagnosed because of limited access to affordable diagnostic tools. Additionally, the development of accessible tools for diagnosing and predicting cognitive impairment has not been extensively discussed in the literature. The objective of this study is to develop a cost-effective and highly accessible machine learning model to predict the risk of cognitive impairment for up to five years before clinical insight. We utilized easily accessible data from the National Alzheimer’s Coordinating Center (NACC) Uniform Data Set (UDS) to train and evaluate various machine learning and deep learning models. A novel algorithm was developed to facilitate the selection of cost-effective models that offer high performance while minimizing development and operational costs. We conducted various assessments, including feature selection, time-series analyses, and external validation of the selected model. Our findings indicated that the Support Vector Machine (SVM) model was preferred over other high-performing neural network models because of its computational efficiency, achieving F2-scores of 0.828 in cross-validation and 0.750 in a generalizability test. Additionally, we found that demographic and historical health data are valuable for early prediction of cognitive impairment. This study demonstrates the potential of developing accessible solutions to predict cognitive impairment early using accurate and efficient machine learning models. Future interventions should consider creating cost-effective assessment tools to support global action plans and reduce the risk of cognitive impairment.
- 22 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 180792-180814
Open AccessOptimizing Machine Learning Models for Accessible Early Cognitive Impairment Prediction: A Novel Cost-Effective Model Selection Algorithm
Cognitive impairment and dementia-related diseases develop several years before moderate or severe deterioration in cognitive function occurs. Nevertheless, most dementia cases, especially in low- and middle-income countries, remain undiagnosed because of limited access to affordable diagnostic tools. Additionally, the development of accessible tools for diagnosing and predicting cognitive impairment has not been extensively discussed in the literature. The objective of this study is to develop a cost-effective and highly accessible machine learning model to predict the risk of cognitive impairment for up to five years before clinical insight. We utilized easily accessible data from the National Alzheimer’s Coordinating Center (NACC) Uniform Data Set (UDS) to train and evaluate various machine learning and deep learning models. A novel algorithm was developed to facilitate the selection of cost-effective models that offer high performance while minimizing development and operational costs. We conducted various assessments, including feature selection, time-series analyses, and external validation of the selected model. Our findings indicated that the Support Vector Machine (SVM) model was preferred over other high-performing neural network models because of its computational efficiency, achieving F2-scores of 0.828 in cross-validation and 0.750 in a generalizability test. Additionally, we found that demographic and historical health data are valuable for early prediction of cognitive impairment. This study demonstrates the potential of developing accessible solutions to predict cognitive impairment early using accurate and efficient machine learning models. Future interventions should consider creating cost-effective assessment tools to support global action plans and reduce the risk of cognitive impairment.
- Open Access
- 31 May 2024
- Published by Society of Visual Informatics in JOIV : International Journal on Informatics Visualization
- Vol. 8 (2) , 1030-1039
Machine Learning-Driven Stroke Prediction Using Independent Dataset
The incidence of stroke cases has witnessed a rapid global rise, affecting not only the elderly but also individuals across all age groups. Accurate prediction of stroke occurrence demands the utilization of extensive data pre-processing techniques. Moreover, the automation of early stroke forecasting is crucial to prevent its onset at the initial stage. In this study, stroke prediction models are evaluated to estimate the likelihood of stroke based on various symptoms such as age, gender, pre-existing medical conditions, and social variables. The machine learning techniques employed include Linear Support Vector Classifier, Extreme Gradient Boosting Classifier, Multilayer Perceptron, Adaptive Boosting Classifier, Bootstrap Aggregating Classifier, and Light Gradient-Boosting Machine. The purpose of this paper is to optimize the hyperparameters of machine learning approaches in developing stroke prediction models. The goal was achieved through a comprehensive comparison of three different sampling techniques for handling imbalanced datasets and evaluating their performance by using various metrics. The most effective model is identified, which is the Adaptive Boosting Classifier utilizing the Tomek Links, with a cross-dataset accuracy of 99% which demonstrated a reliable performance and generalization as evidenced by high cross-validation scores and accuracy on an independent dataset. The next stage of this endeavor entails looking into multiple ways to forecast the development of new dangerous diseases such as breast cancer and skin disorders. In the long run, the aim of subsequent work is to build a powerful toolset that is obtainable to all medical practitioners, allowing for the pre-emptive diagnosis of all potentially hazardous illnesses.
- 31 May 2024
- Published by Society of Visual Informatics in JOIV : International Journal on Informatics Visualization
- Vol. 8 (2) , 1030-1039
Open AccessMachine Learning-Driven Stroke Prediction Using Independent Dataset
The incidence of stroke cases has witnessed a rapid global rise, affecting not only the elderly but also individuals across all age groups. Accurate prediction of stroke occurrence demands the utilization of extensive data pre-processing techniques. Moreover, the automation of early stroke forecasting is crucial to prevent its onset at the initial stage. In this study, stroke prediction models are evaluated to estimate the likelihood of stroke based on various symptoms such as age, gender, pre-existing medical conditions, and social variables. The machine learning techniques employed include Linear Support Vector Classifier, Extreme Gradient Boosting Classifier, Multilayer Perceptron, Adaptive Boosting Classifier, Bootstrap Aggregating Classifier, and Light Gradient-Boosting Machine. The purpose of this paper is to optimize the hyperparameters of machine learning approaches in developing stroke prediction models. The goal was achieved through a comprehensive comparison of three different sampling techniques for handling imbalanced datasets and evaluating their performance by using various metrics. The most effective model is identified, which is the Adaptive Boosting Classifier utilizing the Tomek Links, with a cross-dataset accuracy of 99% which demonstrated a reliable performance and generalization as evidenced by high cross-validation scores and accuracy on an independent dataset. The next stage of this endeavor entails looking into multiple ways to forecast the development of new dangerous diseases such as breast cancer and skin disorders. In the long run, the aim of subsequent work is to build a powerful toolset that is obtainable to all medical practitioners, allowing for the pre-emptive diagnosis of all potentially hazardous illnesses.
- 21 June 2023
- Published by Cosmos Scholars Publishing House in International Journal of Membrane Science and Technology
- Vol. 10 (2) , 1578-1586
Digitizing ECG Signal using 2D Signal Convolution Approach
ECG signal printed on a graph paper has been widely used by medical examiners to analyze diseases related to the heart. Medical practitioners rely on historical records to perform diagnosis. Constantly accessing the ECG printed graph paper manually could be time consuming as there are bulk of graph papers for examination. The proposed work aims to convert the printed ECG graph paper into digitized ECG for remote diagnosis. The ECG printed graph paper undergoes conversion into ECG artifact before transforming as digitized ECG. In the initial phase, patient information in the ECG artifact is preserved by encoding into a QR Code. In phase two, preliminary processing is done on ECG artifact for removal of gridline in the background. Image convolution method is proposed as the process for background gridline removal. Then, morphological image processing is implemented to enhance the ECG artifact. In phase three, segmentation process takes place, in which the ECG artifact is divided into segments for separating the waveforms. In the final phase of ECG digitization, the location of the signal is traced for reshaping the ECG artifact as digitized ECG. The accuracy of the ECG digitization is measured through the heart rate that is calculated using our approach and compared with the one on ECG printed graph paper. The average sum of squared error of the heart rate between the ECG printed graph paper and digitized ECG is 0.005618. The digitized ECG can be useful for medical examiners and practitioners in telemedicine where remote diagnosis may be needed.
- 21 June 2023
- Published by Cosmos Scholars Publishing House in International Journal of Membrane Science and Technology
- Vol. 10 (2) , 1578-1586
Digitizing ECG Signal using 2D Signal Convolution Approach
ECG signal printed on a graph paper has been widely used by medical examiners to analyze diseases related to the heart. Medical practitioners rely on historical records to perform diagnosis. Constantly accessing the ECG printed graph paper manually could be time consuming as there are bulk of graph papers for examination. The proposed work aims to convert the printed ECG graph paper into digitized ECG for remote diagnosis. The ECG printed graph paper undergoes conversion into ECG artifact before transforming as digitized ECG. In the initial phase, patient information in the ECG artifact is preserved by encoding into a QR Code. In phase two, preliminary processing is done on ECG artifact for removal of gridline in the background. Image convolution method is proposed as the process for background gridline removal. Then, morphological image processing is implemented to enhance the ECG artifact. In phase three, segmentation process takes place, in which the ECG artifact is divided into segments for separating the waveforms. In the final phase of ECG digitization, the location of the signal is traced for reshaping the ECG artifact as digitized ECG. The accuracy of the ECG digitization is measured through the heart rate that is calculated using our approach and compared with the one on ECG printed graph paper. The average sum of squared error of the heart rate between the ECG printed graph paper and digitized ECG is 0.005618. The digitized ECG can be useful for medical examiners and practitioners in telemedicine where remote diagnosis may be needed.
- Open Access
- 25 December 2022
- Published by Atlantis Press SARL
- p. 513-523
Lossless ECG Signal Compression Using Non-linear Predictor and ASCII Character Encoding
- 25 December 2022
- Published by Atlantis Press SARL
- p. 513-523
Open AccessLossless ECG Signal Compression Using Non-linear Predictor and ASCII Character Encoding
- Open Access
- 12 September 2022
- Published by F1000 Research Ltd in F1000Research
- Vol. 10, 1052
Job characteristics of a Malaysian bank’s anti-money laundering system and its employees’ job satisfaction
Background: Banks and financial institutions are vulnerable to money laundering (ML) as a result of crime proceeds infiltrating banks in the form of significant cash deposits. Improved financial crime compliance processes and systems enable anti-ML (AML) analysts to devote considerable time and effort to case investigation and process quality work, thereby lowering financial risks by reporting suspicious activity in a timely and effective manner. This study uses Job Characteristics Theory (JCT) to evaluate the AML system through the job satisfaction and motivation of its users. The purpose of this study is to determine how satisfied AML personnel are with their jobs and how motivated they are to work with the system. Methods: This cross-sectional study used JCT to investigate the important elements impacting employee satisfaction with the AML system. The five core dimensions of the job characteristics were measured using a job diagnostic survey. The respondents were employees working in the AML department of a Malaysian bank, and the sample group was chosen using a purposive sampling approach. A total of 100 acceptable replies were gathered and analysed using various statistical approaches. A motivating potential score was generated for each employee based on five main job characteristics. Results: Findings revealed that five core job characteristics, namely, skill diversity, task identity, task importance, autonomy and feedback, positively influence the AML system employees’ job satisfaction. However, skill variety and autonomy are found to be low, which are reflected in the poor motivating potential score. Conclusion: This study examined the characteristics of the AML system and its users’ job satisfaction. Findings revealed that task significance is the most widely recognised characteristic, followed by feedback and task identity. However, there is a lack of skill variety and autonomy, which must be addressed to improve employee satisfaction with the AML system.
- 12 September 2022
- Published by F1000 Research Ltd in F1000Research
- Vol. 10, 1052
Open AccessJob characteristics of a Malaysian bank’s anti-money laundering system and its employees’ job satisfaction
Background: Banks and financial institutions are vulnerable to money laundering (ML) as a result of crime proceeds infiltrating banks in the form of significant cash deposits. Improved financial crime compliance processes and systems enable anti-ML (AML) analysts to devote considerable time and effort to case investigation and process quality work, thereby lowering financial risks by reporting suspicious activity in a timely and effective manner. This study uses Job Characteristics Theory (JCT) to evaluate the AML system through the job satisfaction and motivation of its users. The purpose of this study is to determine how satisfied AML personnel are with their jobs and how motivated they are to work with the system. Methods: This cross-sectional study used JCT to investigate the important elements impacting employee satisfaction with the AML system. The five core dimensions of the job characteristics were measured using a job diagnostic survey. The respondents were employees working in the AML department of a Malaysian bank, and the sample group was chosen using a purposive sampling approach. A total of 100 acceptable replies were gathered and analysed using various statistical approaches. A motivating potential score was generated for each employee based on five main job characteristics. Results: Findings revealed that five core job characteristics, namely, skill diversity, task identity, task importance, autonomy and feedback, positively influence the AML system employees’ job satisfaction. However, skill variety and autonomy are found to be low, which are reflected in the poor motivating potential score. Conclusion: This study examined the characteristics of the AML system and its users’ job satisfaction. Findings revealed that task significance is the most widely recognised characteristic, followed by feedback and task identity. However, there is a lack of skill variety and autonomy, which must be addressed to improve employee satisfaction with the AML system.