Research
Research Interest
Deep Learning
Computer Vision
Publications
Author: Azam, F. B., Carney, R. M., Kariev, S., Nallan, K., Subramanian, M., Sampath, G., ... & Chellappan, S. Classifying stages in the gonotrophic cycle of mosquitoes from images using computer vision techniques. 2023. Scientific Reports (DOI )
Co-Author: Limayem A, Mehta M, Azam F. Evaluation of bactericidal effects of silver hydrosol nanotherapeutics against Enterococcus faecium 1449 drug resistant biofilms. 2023. Frontiers in Cellular and Infection Microbiology. 12 (DOI)
Author: Azam F, Ansari I, Hassan T. Cardiac anomaly detection considering an additive noise and convolutional distortion model of heart sound recordings. 2022. Artificial Intelligence in Medicine. 133 (DOI)
Co-Author: Carney R, Mappes C, Azam F. Integrating global citizen science platforms to enable next-generation surveillance of invasive and vector mosquitoes. 2022. Insects.13(8), 675 (DOI)
Co-Author: Low R, Nelson P, Azam F. Citizen Science-Enabled Tools for the Global Surveillance and Control of Mosquitoes. 2022. Fall Meeting 2022.
--- For additional information, please explore my Google Scholar profile
Research Projects
MosquitoAI
Supervised by Dr. Sriram Chellappan
Professor, Computer and Science Engineering,
University of South FLorida (USF)
Represents a pioneering effort committed to harnessing the power of artificial intelligence for the precise classification of various facets of mosquito biology, encompassing genus, species, instar, and gender.
Applied multiple state-of-the-art deep learning classificaiton models (for example efficientNET, exceptionNET, mobileNet, convNeXt, etc.,) and chose the optimized model through explainable AI techniques (CAM, Grad-CAM, t-SNE, U-map, etc.,).
With advanced AI techniques, MosquitoAI aims to furnish indispensable tools for the surveillance and management of mosquito-related issues, with a particular focus on communities susceptible to mosquito-borne diseases. All the AI models are being hosted through our website, making them accessible for citizen scientists engaged in mosquito monitoring.
Provided the potential to foster community involvement through citizen science platforms, empowering the public to actively partake in mosquito monitoring.
Valvular Disease Classification from Heartsound Database
Supervised by Dr.Taufiq Hasan
Assistant Professor, Biomedical Engineering,
Bangladesh University of Engineering and Technology (BUET)
Addressed the challenge of automatic cardiac abnormality detection in the presence of additive noise and sensor-related issues during cardiac auscultation.
Proposed a method which is to combine linear and logarithmic spectrogram-image features and employ a ResNet classifier, achieving high accuracy, particularly in noisy conditions.
Extensive experiments on multiple datasets to demonstrate the effectiveness of the approach in minimizing the impact of background noise and sensor variability, making it promising for computer-aided cardiac auscultation systems in noisy environments.
Our method offered significant improvements in performance, with impressive metrics such as an AUC of 91.36% and an F-1 score of 84.09%.
Robust Human Authentication using Dynamic ROI Extraction from Dorsal or Palm Hand Vein Images
Supervised by Dr. Mohammed Imamul Hassan Bhuiyan
Professor and Head, Biomedical Engineering,
Bangladesh University of Engineering and Technology (BUET)
Employed ResNet50 model training to identify the image's region of interest.
Conducted training on two demanding databases and implemented Domain Adaptation through Domain Adversarial Neural Network (DANN) to enhance domain feature independence.
Developed a user-friendly MATLAB GUI interface to streamline the feature extraction process.