Projects

The motive behind this project is to propose the sophisticated supervised machine learning model in the prediction and classification of AD in elder people. For that, we conducted different experiments on open access brain image information including demographic MRI data of 373 scan sessions of 150 patients. In the first two works, we applied single ML models called support vectors and pruned decision trees for prediction of dementia on the same dataset. In the first experiment with SVM, we achieved 70% of prediction accuracy of late-stage dementia. Classification of true dementia subjects (precision) is calculated as 75%. Similarly, in the second experiment with J48 pruned decision trees, the accuracy was improved to the value of 88.73%. Classification of true dementia cases with this model was comprehensively done and achieved 92.4% of precision. To enhance this work, rather than single modelling we employed multi-modelling approaches. In the comparative analysis of machine learning study, we applied the feature reduction technique called principal component analysis. This approach identifies the high correlated features in the dataset that closely associated dementia type. By doing the simultaneous application of three models such as KNN, LR, and SVM, it has been possible to identify an ideal model for classification of dementia subjects. When compared with support vectors, KNN and LR models comprehensively classified AD subjects with 97.6% and 98.3% of accuracy respectively. These values are relatively higher than the previous experiments. However, because of the AD severity in older adults, it should be mandatory to not leave true AD positives. For the classification of true AD subjects among total subjects, we enhanced the model accuracy by introducing three independent experiments. In this work, we incorporated two new models called Naïve Bayes and Artificial Neural Networks along support vectors and KNN. In the first experiment, models were independently developed with manual feature selection. The experimental outcome suggested that KNN is the optimal model solution because of 91.32% of classification accuracy. In the second experiment, the same models tested with limited features (with high correlation). SVM was produced high 96.12% of classification accuracy and NB produced 98.21% classification rate of true AD subjects. Ultimately, in the third experiment, we mixed these four models and created a new model called hybrid type modelling. Hybrid model performance is validated AU-ROC curve value which is 0.991 (i.e., 99.1% of classification accuracy) has achieved. All these experimental results suggested that the ensemble modelling approach with wrapping is an optimal solution in the classification of AD subjects.

The occupational sector of maritime workers is distinguished from all other occupations by its own unique economic and social characteristics, but also and especially by sometimes unique issues related to occupational safety. In fact, maritime workers may be subject to situations of significant fatigue and work stress, are exposed to noise and vibration, radiation, biological and chemical agents, night work, sudden changes in microclimate, and risk from manual handling of loads. Moreover, seafarers experience a greater tendency aging than the general population due to the wearing nature of some activities performed and voluptuous habits (smoking, alcohol).

Protecting the health of seafarers is both an important aspect from an ethical point of view, depending on the varied occupational hazards that characterize work at sea, and an objective requirement depending on the safety of navigation and the protection of the substantial economic interests associated with it.

This is the background to the activity developed by TelePharmaTec Srl in collaboration with the International Radio Medical Center (C.I.R.M.) service company CIRM SERVIZI Srl, to monitor work-related stress on board ships at sea. This is in the function of protecting sea workers and increasing the safety of navigation.

This project aims to develop a data observatory of seafarers to improve health status by monitoring communicable and non-communicable diseases of seafarers by monitoring their health and disease trends, creating a standard tool (questionnaires), and conducting a risk assessment.  I contributed to the data integration process by collecting and integrating data from various sources. This involved synchronizing the data and ensuring that the information were regularly updated to provide a complete overview. A significant part of my contribution has been in developing customized dashboards for different groups of users. These dashboards provided real-time information and industry-specific tools their needs, improving their decision-making skills. I played a role in implementing automated systems that monitored and continuously reported compliance with maritime regulations. This helped ensure that the maritime industry operated in compliance with the necessary regulations. My contribution has involved the integration of data visualizations, such as graphs and maps, to provide useful information for users. These visualizations helped users understand better trends and make informed decisions. Throughout the project, I actively participated in the collaborative framework, which has involved structured phases and regular communication. We used project platforms management, video conferencing, version control and dedicated communication channels for ensure effective collaboration.

The purpose of “MARINE DOCTOR” is to simplify the medical assistance of seafarers. The developed software is user-friendly, simple, fast, and compatible with any computer. The main function of the system is to register and store seafarers' and doctors' details and guide the user to send a medical request with a simple form of a questionnaire. In the development of the "MARINE DOCTOR" software, I played a pivotal role in creating a user-friendly and efficient system to streamline medical assistance for seafarers. My primary contributions include the design and implementation of the front-end graphical user interface (GUI) and the integration of robust database connections. These contributions ensure that the software is intuitive, responsive, and compatible with any computer system.  Front-End GUI  a clean, simple, and intuitive graphical user interface using Tkinter, a standard Python library for GUI development. Designed and set up a relational database to efficiently store and manage seafarers' and doctors' details. Created a structured schema to organize the data effectively, enhancing the system's ability to quickly and accurately retrieve information as needed. Developed the core functionality to facilitate the registration and management of seafarers' and doctors' details. Implemented a streamlined process for users to send medical requests via a simple questionnaire form, reducing the complexity and time required for submitting requests. Ensured compatibility with various computer systems, focusing on cross-platform functionality to cater to a diverse user base. 

MarineDerma is a lightweight, AI-powered desktop application designed to assist in the early identification of skin lesions and infectious skin conditions. Trained on a diverse dataset across 14 categories — including melanoma, carcinomas, viral infections (like monkeypox), and healthy skin — MarineDerma uses a MobileNetV3-based model to deliver accurate predictions with up to 91% internal test accuracy.

Built for offline use, the tool is available as a standalone .exe file, allowing healthcare teams and researchers to use it on board ships, in remote areas, or in low-connectivity environments.