Scholar
Akara Supratak, PhD
Google Scholar ID: r1KTiH4AAAAJ
Assistant Professor of Computer Science, Mahidol University
Citations & Impact
Citations
2,259
H-index
11
i10-index
11
Publications
20
Co-authors
0
Contact
Publications
Browse publications on Google Scholar (top-right) ↗
Resume (English only)
Academic Achievements
- Developed a series of models for automatic sleep stage scoring using raw, single-channel EEG, including DeepSleepNet and TinySleepNet, and introduced a method for quantifying the transferability of sleep stage scoring models; Aiming to develop deep learning models capable of detecting brain conditions directly from CT scans; Designed MosquitoSong+, a new deep learning model that identifies the species and sex of mosquitoes robustly against environmental noise; Developed a platform to automatically extract blink dynamics from high-frame-rate videos; Developing and testing a federated learning platform for collaborative learning across multiple institutions while preserving data privacy; Collaborating on an end-to-end deep learning model for asteroid detection.
Research Experience
- Assistant Professor at the Faculty of ICT, Mahidol University, 2022-current; Instructor at the Faculty of ICT, Mahidol University, 2018-2022. Engaged in multiple research projects including advancing sleep stage scoring with deep learning, developing deep learning models for brain CT classification, creating a noise-robust mosquito species and sex classification model, quantifying and tracking objective blinking parameters, establishing a federated learning platform for medical imaging analysis, and working on deep learning for asteroid detection in sky exposures.
Education
- Doctor of Philosophy (Ph.D.), Computing Research, Imperial College London, 2013-2018.
Background
- My research focuses on biomedical informatics and deep learning, especially training models with limited data and resources for medical image and signal analysis.
Co-authors