Browsing by Author "Abdal, M.N."

Now showing 1 - 3 of 3

A Lightweight-CNN Model for Efficient Lung Cancer Detection and Grad-CAM Visualization
(Institute of Electrical and Electronics Engineers Inc., 2023) Tasmim,; Bakchy, S.C.,; Peyal, H.I.,; Islam, M.I.,; Yeamin, G.K.; Miraz, S.,; Abdal, M.N.
The lungs' abnormal cell growth leads to the development of lung cancer. Early cancer identification could make treatment easier, potentially saving millions of lives annually. This study's main goal is to more rapidly and effectively classify various types of lung cancer by employing a lightweight, computationally efficient convolutional neural network (CNN) model to categorize three different types of lung cancer. With an outstanding validation accuracy of 99.48%, the suggested model surpasses the achievements of previous works. The 15,000 CT scan images in our dataset include three different forms of lung cancer. The proposed model performs exceptionally well, as evidenced by its astounding precision, recall, and F1-score, all above 99%, and by its flawless Area Under Curve (AUC) score of 100%. The proposed model has fewer parameters than the existing transfer learning models. Gradient Weighted Class Activation Mapping (Grad-CAM) was used to create class activation maps, which were then used to create a heatmap to display the classification zone.
An Interpretable Framework for Predicting Type 2 Diabetes using ML and Explainable AI
(Institute of Electrical and Electronics Engineers Inc., 2023-12-13) Ahmed, K.F.; Uz Zaman, M.S.; Hossain, A.; Rahman Ratul, M.T.; Abdal, M.N.; Islam, M.I.
Diabetes Mellitus is an incurable disease and stands as a major universe cause of destruction. With the universal prevalence of diabetes increasing rapidly, accurate detection and identification of the disease have become crucial. In this paper, we present an explainable AI-based diabetes indicative methodology that is carefully constructed, effective, and, most importantly, interpretable. Using two real-world diabetes datasets, the three most well-known machine learning classifiers in the literature - Random Forest (RF), Decision Tree (DT), and XGBoost - were subjected with quantitative assessments. After training and assessment of all classification models, the suggested method achieved the best results in the XGBoost classifier for the Sylhet dataset with 99.4% accuracy and the Pima Indian dataset with 92.59% accuracy. The datasets were normalized across observations to provide standardized values for improved analytical applicability and comparability. Explainable AI has been implemented for the XGBoost machine learning model by generating both global and local explanations using Shapley additive explanations (SHAP) and Local interpretable model-agnostic explanations (LIME). The elements that contribute to diabetes are explained and shown in graphs to help medical professionals make decisions about clinical diagnosis and treatment options.
Detection of Primary Open Angle Glaucoma Based on Deep CNN Using Fundus Images
(Institute of Electrical and Electronics Engineers Inc., 2023-12-13) Ratul, M.T.R.; Afroge, S.; Peyal, H.I.; Zafrin, D.; Ahmed, K.F.; Abdal, M.N.
The early and accurate diagnosis of glaucoma, a primary cause of permanent blindness, is critical for efficient treatment and prevention of vision loss. Although the exact causes of glaucoma are not yet fully understood, it is thought to be a result of several factors, including raised pressure inside the eye and decreased blood supply to the optic nerve. We have developed a convolutional neural network model for accurate detection of glaucoma. Methods based on deep learning have been effective at classifying diseases in retinal fundus images, facilitating in the evaluation of the growing number of images. The goal of this work is to create and train a unique deep CNN model that makes use of the connections between related eye-fundus tasks and metrics used to identify glaucoma. We have meticulously selected two distinct datasets to underpin this research endeavor: the ACRIMA dataset and the LAG dataset. Notably, our model attains a remarkable accuracy score of 99.29% on the ACRIMA dataset and an equally commendable accuracy score of 97.22% on the LAG dataset. This performance eclipses that of the majority of contemporary deep CNN models, underscoring the prowess and sophistication of our approach