My research interests lie in applying machine learning and deep learning techniques to real-world healthcare challenges.
I am currently working on Non-Rigid Structure from Motion (NRSfM), 3D reconstruction, and deformable modeling for medical imaging,
with a particular focus on Neural Radiance Fields (NeRF) and non-rigid registration. My research aims to develop high-fidelity and robust
methods for reconstruction and registration in minimally invasive surgery (MIS). Previously, I worked on Human Activity Recognition (HAR), with an emphasis on fall detection for elder people and monitoring.
Evaluation autonomous on-wrist wearable devices using different trained machine learning models, where a
single dataset is used for training and multiple independent datasets are used for validation.
This paper proposes a hybrid fall detection method combining personalized and generalized models. The approach improves robustness and accuracy across different users compared to single-model method
This paper evaluates a fall detection system based on improved feature extraction and data
balancing techniques to enhance classification performance..
Propose the low-cost computation fall detection system on wearable accelerometer.
A Stochastic Programming Model for Decision-Making Concerning Medical Supply Location and Allocation in Disaster Management
Approach Samad Barri Khojasteh,
Irfan Macit ,
Cambridge, 2017
project page
/
paper
Stochastic programming model as a solution for optimizing the problem of locating and
allocating medical supplies.