Dr. C. P. SHIRLEY
@karunya.edu
Assistant Professor , Computer Science and Engineering
Karunya Institute of Technology and Sciences
RESEARCH INTERESTS
Artificial Intelligence, Machine Learning, Deep Learning
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
K.G. Revathi, C.P. Shirley, S. Sreethar, and Ezhilarasi P
Elsevier BV
Dinesh G, W. Deva Priya, C. P. Shirley, and T. Vignesh
Springer Science and Business Media LLC
S. Soundararajan, C. P. Shirley, Balasubbareddy Mallala, and K. Padmanaban
Springer Science and Business Media LLC
Vidhya K, Nagarajan B, Jenefa A, Catherine Joy R, C. P. Shirley, and Joel J
IEEE
Timely identification of cervical cancer is essential for effective therapy and enhanced patient outcomes. Conventional imaging methods, although fundamental, frequently inade-quately represent the Detailed characteristics of cancer pathology because they depend on isolated data sources. MedFuseN et tackles these difficulties by presenting an advanced architecture that integrates multi-modal data to improve diagnostic accuracy and dependability significantly. This advanced model employs high-resolution medical imaging and incorporates patient-specific clinical data, establishing a comprehensive analytical foundation for diagnosis. Our dataset includes 4,049 annotated cervical cell images across a spectrum from normal to malignant, each enriched with detailed clinical parameters. MedFuseNet utilizes a hybrid architecture that integrates CNNs for image data and RNNs for sequential clinical data, facilitating a thorough analysis of various data sources. The methodological integration enables MedFuseN et to surpass conventional single-source models, with an accuracy of 98.5 %, alongside significant enhancements in pre-cision (97.9 %) and recall (98.1 %). These substantial diagnostic improvements highlight the promise of multi-modal data fusion in medical imaging, paving the way for creating more sophisticated, AI -driven diagnostic instruments that may revolutionize early cancer diagnosis and treatment approaches.
Immanuel John Raja, S. V. Evangelin Sonia, C. P. Shirley, and I. Titus
Springer Science and Business Media LLC
M. Maheswari, Mohamed Uvaze Ahamed Ayoobkhan, C. P. Shirley, and T. R. Vijaya Lakshmi
Springer Science and Business Media LLC
C. P. Shirley, Depaa RA B, A. Priya, R. Sarala, Rajdeep Singh Solanki, Malini K. V, and Ch. Venkatakrishna Reddy
The Institute of Electronics Engineers of Korea
C. P. Shirley, J Immanuel John Raja, S. V. Evangelin Sonia, and I. Titus
Springer Science and Business Media LLC
Elakkiya Elango, Gnanasankaran Natarajan, Ahamed Lebbe Hanees, and Shirley Chellathurai Pon Anna Bai
IGI Global
Organizations are quickly realizing the transformative possibilities of digital twins and artificial intelligence (AI) in this era of fast technical advancement. This chapter provides a brief synopsis of “The Roadmap to AI and Digital Twin Adoption,” a comprehensive resource that delves into the key elements and techniques necessary for the successful integration of AI and digital twins across a range of sectors. This roadmap explores the mutually beneficial relationship between artificial intelligence (AI) and digital twins, emphasizing how each may enhance overall performance, decision-making, and operational efficiency. It covers the fundamental concepts of artificial intelligence (AI), such as natural language processing, machine learning, and deep learning, and how important they are in relation to digital twins. The guide's emphasis extends to the practical use of AI and digital twins, offering guidance on data collection and management, model training, and algorithm choice.
Gracious S and Shirley C P
IEEE
Heart disease is still the world’s top cause of death, and becoming older is a major risk factor. Age-related risk for heart disease can be decreased by using early detection and prevention strategies. This research study presents a novel machine learning-based method for estimating an individual’s risk of heart disease based on age-related variables and their medical history. This study analyzes the efficacy of ensemble learning methods, namely XGBoost and Random Forest models, in predicting the risk of heart disease by utilizing a tabular dataset that includes data on individuals' health records from various age groups. By means of thorough experimentation and model validation, the proposed model classifies heart disease with 93.26% accuracy. This research adds to the advancement of predictive modeling for age-related heart disease and holds promise for personalized healthcare interventions aimed at improving patient outcomes and lowering healthcare costs. The research findings not only highlight the potential of machine learning in healthcare but also highlight the significance of early intervention strategies tailored to age-related factors.
Shirley C P and Sujith P
IEEE
In recent years, piracy has become a common practice in the film and gaming industry. Previously, films were pirated, copied and distributed using compact disks (CDs), but now pirated content is shared online. In the gaming industry, competitors and fans often copy textures, characters, maps, and other content from each other. The proposed method aims to address the issue of piracy that occurs within the industry by proposing a security framework that uses cryptography for encryption and digital watermarking techniques for identification. Content creators usually have to share their work with other developers or creators in a work environment, which makes the content vulnerable to piracy. The aim of the proposed work is to put the created content into this security framework, encrypt it, and embed an invisible watermark specific to that user when sharing it with others. If any other creator needs that content, they can request it through the license server, and it will be decrypted using the keys from the server. The proposed method ensures that leaked content will have the watermark, which can help identify the person who leaked it, be it a city map in a game or a scene from a movie. This article will explain the methodologies used in this approach in detail.
Shirley C P and Absin S J
IEEE
Facial expression recognition is becoming a core part of human-computer interaction and emotion detecting systems. However, the real-time scenarios such as the widespread use of face masks presents a big challenge to traditional facial recognition systems. This work proposes a new approach by integrating machine learning concepts, particularly in neural networks and feature fusion. Our main goal is to develop an accurate facial recognition system that can recognize expressions even when the face is partially visible. To achieve this process, we will take a diverse dataset containing images of faces with variable occlusion. Using Data pre-processing techniques, it standardized the training and evaluation procedure. The obtained results conclude the model's ability to accurately recognize facial expressions even in the presence of occluded faces. To achieve this objective, this research work has proposed a Squeeze and Excitation Network (SE-Net) to enhance the relationship in convolutional feature maps. The proposed feature fusion technique combines the local features and global features. Moreover, it creates opportunities for enabling more consideration into scientific research and the synthesis of data from various inputs by eventually validating more complex situations.
Kavin S and Shirley C P
IEEE
To guarantee patient safety and regulatory compliance, healthcare providers must accurately enter batch numbers and expiration dates from drug packaging. But manual entry is prone to mistakes, which could have unfavourable effects. This project suggests an automated method for extracting batch numbers and expiration dates from the back of pharmaceutical packaging by using optical character recognition (OCR) technology. In order to improve OCR accuracy, the research makes use of preprocessing techniques and a broad dataset of photos from medical packaging. The retrieved text is analysed to determine the batch number and expiration date, and data correctness is ensured by validation procedures. The end product solution reduces errors in healthcare data management by offering an intuitive user interface for effective data extraction. By means of thorough inspection and evaluation, The experiment shows how the OCR-based system can reliably and effectively extract important information from pharmaceutical packaging. Through the reduction of data entry errors and the facilitation of error-free information retrieval from drug packaging, this effort seeks to improve patient safety and streamline healthcare procedures.
S. Alvin Jesuraj, S. V. Evangelin Sonia, and C. P. Shirley
IEEE
This research investigates the application of DenseNet-201, a deep convolutional neural network architecture, in the RSNA-MICCAI Brain Tumor Radiogenomic Classification aimed at predicting the genetic subtype of glioblastoma using MRI imaging data. This study demonstrates the effectiveness of DenseNet-201 in accurately classifying glioblastoma cases based on MGMT promoter methylation status, a critical biomarker influencing treatment outcomes. Through comprehensive experimental evaluations, including training, validation, and testing phases, DenseNet-201 exhibits robust performance metrics such as high accuracy, precision, recall, F1-score, and AUC-ROC values. These results highlight the model's ability to effectively distinguish between MGMT promoter methylation-positive and negative glioblastoma cases, offering valuable support for clinical decision-making in treatment planning and prognosis assessment. Leveraging deep learning techniques and MRI imaging data, DenseNet-201 holds promise as a powerful tool for enhancing the understanding of glioblastoma genetics and guiding personalized therapeutic interventions, ultimately contributing to improved patient outcomes in brain cancer management.
Shirley C P, Brijit Benny, Vidhya K, and Berin Jeba Jingle I
IEEE
Agriculture is the activity of cultivating natural resources that humans utilize for survival and economic benefit. It has evolved into one of the most significant industries in the day-to-day existence of all living organisms, providing access to food, a means of subsistence, etc. In reality, a country's agricultural output may have a direct or indirect impact on its food security and overall health. Despite the widespread adoption of these agricultural methods, various obstacles, including soil fertility and climate, can negatively impact the development and output of agricultural goods. As every living thing on Earth depends on agriculture for its existence, there is a constant need to find ways to overcome the problems faced. It is a well-acknowledged reality that contemporary issues demand modern solutions. Thus, it is critical to use developing technology to tackle the challenges that exist. The quality of life has been greatly enhanced by modern technologies. This concept has led to the introduction of several contemporary inventions and technologies that enhance the standard of life and advance the welfare of every individual. The demand for new technologies has been growing over time, and one promising technology with the potential to improve living standards and address long-standing issues in the agricultural sector is digital twin technology. It has made it possible to reduce labor and energy waste, among other things. Examining how digital twins might improve agricultural productivity and promote sustainability is part of the process of understanding digital twins and their place in the agricultural industry. The prerequisites and implementation strategies for encouraging the use of digital twins are covered. Since this is a new technology, a careful analysis of its benefits and limitations is necessary to properly address identified issues.
Venkatesan R, Brijit Benny, Shirley C P, and Berin Jeba Jingle I
IEEE
As the virtual and real worlds continue to collide, digital twins may become important tools for finding solutions to some of the most serious challenges facing the sustainability movement. These challenges include making effective use of resources, adjusting to the effects of climate change, and ensuring that everyone has access to a high quality of life. The first steps have been taken on the road to making widespread use of digital twins. It is imperative that their connectedness should be secured if their potential is to be fully utilized. This presents a number of challenging situations. The digital transition may make it possible to achieve several of the United Nations' sustainable development objectives. Data and technology should assist the user to make appropriate judgements that are better suited to their circumstances. It’s feasible that the computer model can see into the future and predict the user about the consequence. This may be used to make predictions about the future, analyze the present, and analyze how a decision will play out in the future. These features may have an effect on a number of different sectors including commerce. The use of dynamic modelling may be beneficial for improving efficiency in the management of supply chains, transportation, and crowds.
J. Immanuel Johnraja, P. Getzi Jeba Leelipushpam, C. P. Shirley, and P. Joyce Beryl Princess
Springer Nature Switzerland
Shirley C P, Thanga Helina S, Berin Jeba Jingle I, Saran P, and Absin S J
IEEE
Blockchain technology's decentralized and transparent transaction platforms have changed a number of sectors. However, harmful activities like fraud and money laundering are also drawn to blockchain due to its distributed and unchangeable nature. The distinctive features of blockchain transactions frequently prove to be too much for conventional fraud detection techniques to handle. In order to identify fraud in blockchain transactions, this work suggests "Secure Sentinel," a machine learning-based method based on the Isolation Forest concept. By separating anomalies in a binary tree structure, the Isolation Forest algorithm is highly effective in detecting abnormalities inside datasets. By utilizing this approach, our system is able to identify transactions on the blockchain that are suspicious or stray from the norm. The Isolation Forest model will be trained using attributes such transaction amount, frequency, source/destination addresses, and transaction time. The project's goal is to improve blockchain networks' security and integrity by offering a real-time fraud detection system. Users, miners, and regulators are among the blockchain stakeholders who can take proactive steps to reduce risks and preserve system confidence by automatically identifying potentially fraudulent transactions. We will use past blockchain data to test our model's performance and determine how well it detects different kinds of fraudulent activity in terms of accuracy, precision, recall, and efficiency. By using machine learning to protect blockchain ecosystems from new dangers, Secure Sentinel is a step in the right direction toward advancing the acceptance and sustainability of blockchain technology.
Brijit Benny, Shirley C P, Berin Jeba Jingle I, Vidhya K, and Elakkiya E
IEEE
Digital twins and Internet of Things (IoT) technology have revolutionized modern farming by enabling creative techniques of monitoring and controlling crop growth and well-being. This article investigates the potential applications of Internet of Things (IoT) sensors and digital twin models for agriculture to provide farmers with analytical projections, assistance in making decisions, and continuous surveillance. Digital twin models are simulated representations of the real agricultural ecosystem that use data from multiple sources to correctly represent the dynamic interplay among crop nutritional status and conditions in the environment. The Internet of Things (IoT) and the use of digital twins have revolutionized the modern agricultural sector by enabling the monitoring and management of farms in order to regulate crop health and growth. The current research looks into the use of IoT sensors and digital twin technologies in agriculture. Farmers can obtain comprehensive data on agricultural growth conditions by utilizing Internet of Things (IoT) sensors. Sensors such as these collect data on solar exposure, pH, temperature, humidity, and moisture levels in the soil. Digital twin designs, computational depictions of the real agriculture atmosphere, are being developed using this information sources to replicate the dynamic interaction involving the physiology of crops and external factors.
Shirley C P, Berin Jeba Jingle I, Abisha M B, Venkadesan R, and Absin S J
IEEE
Healthcare decision-making is a complex and crucial process that must take into account changing patient situations and medical trends. It is frequently difficult for conventional decision support systems to adjust to these shifting circumstances. In this work, we present a novel method for creating adaptive healthcare decision support systems by combining time series forecasting and reinforcement learning (RL). The main goal is to create a system that can use reinforcement learning (RL) to develop optimal policies for making decisions by utilizing time series data of patients from the past. Our system uses RL in conjunction with time series forecasting methods like LSTM (Long Short-Term Memory) networks and ARIMA (Auto Regressive Integrated Moving Average) to estimate future patient outcomes and healthcare resource requirements. There are several different components in the system architecture. The time series forecasting models are trained using historical healthcare data, which includes patient demographics, medical histories, and treatment outcomes. Forecasts for important healthcare parameters like patient recovery rates, disease progression, and resource consumption are produced by these models. According to the anticipated results, RL algorithms are used to learn decision policies. While interacting with a simulated environment that represents patient care scenarios, the RL agent aims to maximize long-term rewards, which may include reduced expenses, better patient outcomes, or more efficient use of available resources. Using actual healthcare datasets, the adaptability and efficacy of suggested approach will be assessed. The efficacy of RL-based decision policies, accuracy of projected results, and the system's adaptability to shifting healthcare dynamics are all examples of performance measures.
Shirley C P, Berin Jeba Jingle I, Abisha M B, Venkatesan R, Yashvanth Ram R V, and Elakkiya Elango
IEEE
The proliferation of deepfake technology poses a significant challenge to the authenticity of digital content. This research explores the application of multimodal fusion techniques to enhance deepfake detection accuracy. By combining visual and audio features, the proposed method leverages the complementary nature of different data types to detect discrepancies introduced by deepfake manipulation. An attention mechanism is incorporated to focus on salient regions within each modality, further improving detection accuracy. Convolutional Neural Networks (CNNs) and Mel-Frequency Cepstral Coefficients (MFCCs) are employed for feature extraction, followed by feature fusion for deepfake detection. This approach demonstrates the effectiveness of multimodal fusion in combating the evolving threat of deepfake technology. By advancing deepfake detection techniques, this research contributes to safeguarding the integrity of digital content and preserving trust in media.
S.V. Evangelin Sonia, C Beulah Christalin Latha, R. Venkatesan, G.Naveen Sundar, C.P. Shirley, and E Harris Stuart
IEEE
Wearable motion capture technology has advanced significantly, enabling detailed and accurate tracking of human movements in three-dimensional space. This paper introduces FocusRec RNN, an innovative Recurrent Neural Network (RNN) architecture designed to enhance the reconstruction and prediction of 3D human poses using wearable IMU sensors. By incorporating a focus mechanism that dynamically prioritizes critical time-steps and body segments, FocusRec RNN improves upon traditional methods in terms of accuracy and robustness. Evaluation on datasets such as Human3.6M and the CMU Motion Capture Database demonstrates that FocusRec RNN achieves lower Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), alongside higher Percentage of Correct Keypoints (PCK) compared to traditional RNNs, LSTM networks, and CNNbased approaches. These improvements highlight FocusRec RNN’s potential for applications in sports performance analysis, medical diagnostics, rehabilitation, and interactive entertainment.