Heshan Chandeepa Pathmakumara

2025 October

A CNN Approach for Accurate Stroke Diagnosis Using Brain Computed Tomography Imaging

Our paper, "A CNN Approach for Accurate Stroke Diagnosis Using Brain Computed Tomography Imaging", co-authored by me and my supervisor Mr. Gayan Perera, has been published in the IEEE 2025 MIUCC Conference Proceedings.

This study introduces a CNN-based technique for detecting strokes from brain CT (computed tomography) images using a dataset of 2,501 scans, including both normal and stroke cases. The proposed model achieved a validation accuracy of 92.93%, demonstrating its high capability to distinguish between normal and stroke images. The evaluation metrics — including ROC curves, accuracy, recall, and F1-score — confirmed the model’s potential as a reliable clinical decision-support tool.

2025 October

Explainable AI for Glaucoma Detection Published in medRxiv

Our paper, "Explainable Deep Learning for Glaucoma Detection: A DenseNet121-Based Classification with Grad-CAM Visualization", co-authored by me and my supervisor Mr. Gayan Perera, has been published as a preprint in medRxiv!

This research presents an explainable AI system that detects glaucoma from retinal fundus images with 90.16% accuracy, using Grad-CAM visualization to show which areas of the eye the model focuses on - making AI decisions transparent and clinically interpretable.

2025 May

Research Paper Published in IEEE Xplore

Our paper, "Advanced Deep Learning Techniques for Lung Sound Classification: Binary, Multi-Class, and Ensemble Approach", co-authored by me and my supervisor Gayan Perera, has been published in IEEE Xplore!

This research presents deep learning methods to improve lung sound classification for respiratory disease detection.

2025 May

Final Year Project Poster Presentation

I successfully defended my final year project, " PulmoSense AI: A Deep Learning Based Lung Sound Classification System", at the poster presentation event.

It was a rewarding experience sharing my work on pulmonary disease detection using deep learning.

2025 May

Thesis Submitted

I submitted my final year thesis titled "PulmoSense AI: A Deep Learning Based Lung Sound Classification System" to the University of Plymouth, under the supervision of Mr. Gayan Perera, for my BSc (Hons) in Data Science.

2025 March

Paper Presentation at CSECS 2025

Me and my supervisor, Gayan Perera, presented our paper, "Advanced Deep Learning Techniques for Lung Sound Classification: Binary, Multi-Class, and Ensemble Approach" at the 7th International Conference on Software Engineering and Computer Science, Xi'an Jiaotong-Liverpool University, China!

Our research introduces deep learning techniques to enhance lung sound classification accuracy.

2024 October

Paper Presentation at ICACT 2024

I presented & published my research, "Machine Learning Techniques for Predicting Brain Stroke Risk: Addressing Data Imbalance" at ICACT 2024!

This work explores how machine learning can tackle data imbalance challenges to improve stroke risk prediction.