PA
About me
A freshly graduate computer science student who passionate in expanding sets of knowledge in the world. Currently, I'm working on federated learning on medical image.
Work Experience
Research Assistant
2024 - Now
Faculty of Information and Communication Technology, Mahidol University
See more detail in Research Project.
Student Research Assistant
2021 - 2024
Faculty of Information and Communication Technology, Mahidol University
Worked on Low-Budget Deep Active Learning on Eyelid Localization, Noise-Robust Mosquito Wingbeat Sound Event Detection, Noise-Robust Mosquito Wingbeat Sound Event Detection, Foot Pad Dermatitis (FPD) Severity Identifier. See more detail in Research Project.
Backend Developer Internship
2020 - 2021
BotNoi Group
Implemented a general dashboard for the company's chat bots showing statistical information, such as the amount of active users, the amount of messages, user retention rate over a specific time period.
Research Project
Low-Budget Deep Active Learning on Eyelid Localization
Faculty of Information and Communication Technology, Mahidol University
Low-Budget Deep Active Learning
Keypoint Localization
Performed experiments on high frame rate videos of patients with abnormal blinking. My work emphasized on implement active learning algorithms in order to select representative data and aimed to reduce data required to be annotated by ophthalmologists.
Noise-Robust Mosquito Wingbeat Sound Event Detection
Faculty of ICT, Mahidol University and BSCC, University of Bremen
Polyphonic Sound Event Detection
Mosquito Classification
Noise-robust Sound Event Detection
Utilized mosquito wing beat sound to conduct experiments on a designed and implemented noise-robust sound event detection model. The model was trained on environmental noise, in which some portion had been overlayed realistically with mosquito’s wing beat sound, then predicted when and which species and sex of mosquitoes were present in the given audio.
Normal-Pressure Hydrocephalus (NPH) Detection
Faculty of Information and Communication Technology, Mahidol University
Image Segmentaion
Weak Supervised Learning
Experimented on CT scans using segmentation and classification models to detect the brain disease. Designed a pipeline that can be trained using weakly supervised data, which does not require clinical annotation. The designed pipeline has more sensitivity than radiologist and thus could be used as a screening tool which reduce time for the physicians and the patients.
Foot Pad Dermatitis (FPD) Severity Identifier
Faculty of Information and Communication Technology, Mahidol University
Image Segmentaion
Image Classification
Designed a pipeline that receives duck feet images and identifies the severity of FPD in each foot. Maximized the accuracy by using ensemble technique and transfer learning.
Publication
Medical Image Processing
Normal Pressure Hydrocephalus
Weak Supervision
NPH Classification
Dittapong Songsaeng, Poonsuta Nava-apisak, Jittsupa Wongsripuemtet, Siripra Kingchan, Phuriwat Angkoondittaphong, Phattaranan Phawaphutanon, and Akara Supratak
Published in Diagnostics.
Medical Image Processing
Normal Pressure Hydrocephalus
Weak Supervision
NPH Classification
Akara Supratak, Siripra Kingchan, Phuriwat Angkoondittaphong, Poonsuta Nava-apisak, Jittsupa Wongsripuemtet, Thanapon Noraset, Worapan Kusakunniran, Peter Haddawy, and Dittapong Songsaeng
To be published in Artificial Intelligence in Medicine.
Active Learning
Keypoint Detection
High-Frame-Rate Videos
Blink Analysis
Akara Supratak, Phuriwat Angkoondittaphong, Napahatai Sittirit, Watsaporn Pornwatanacharoen, Pimkwan Jaru-ampornpan, and Thanapon Noraset
To be published in Applied Soft Computing Journal.
Presentation
Representation Learning
Hierachy Aware Knowledge Base Embedding
Phuriwat Angkoondittaphong, ReaLearn Lab
An attempt for apply Embed2Reason (E2R), which is designed for embed knowledge base ALC into vector space, on RiceDO and TreatO which are EL Ontologies.
Polyphonic Sound Event Detection
Mosquito Classification
Noise-robust Sound Event Detection
Phuriwat Angkoondittaphong, Napahatai Sittirit, Danaidech Ardsamai, Akara Supratak
More detail in Noise-Robust Mosquito Wingbeat Sound Event Detection in Research Project section.
Education
Bachelor of Science in Information and Communication Technology
2020 - 2024
Faculty of Information and Communication Technology, Mahidol University
Major in Computer Science (First Class Honours)
Advisor: Asst. Prof. Dr. Akara Supratak
Skills
Python
TensorFlow
Pytorch
Deep Learning
Sound Event Detection
Image Classification
Image Segmentation
Deep Active Learning