APEC-Australia Women in Research
Dr. Thuy Thi Nguyen
Host Institution: University of Technology Sydney
Project Title: Research into deep learning algorithms to detect lesions in endoscopy images
This project aims to conduct research on an AI-assisted system to detect and identify lesions in high-risk endoscopy images. We focus on proposing a novel algorithm that can achieve state-of-the-art results in the prediction of lesions. A survey will be done on advanced machine learning models, especially recent deep learning algorithms. Then, a deep neural network architecture aiming at efficient computing will be developed. The algorithms will be trained and tested on benchmark datasets and further
improved on our newly created large colonoscopy dataset (available from our ongoing project) to obtain state-of-the-art performance. The algorithms can be embedded in Computer-Aided Diagnosis (CAD) systems, as an AI assisting tool (AI “doctor”) for a clinician to make a diagnosis. This can be applied in clinics to increase lesion detection rate, especially for adenomas and high-risk lesions, as well as to serve an increasing number of patients while maintaining healthcare quality. This is particularly needed in developing nations with overload patients such as Vietnam.
Reasons for applying for fellowship:
The research fellowship from APEC-Australia Women in Research will help me to advance my knowledge in the field, to pave the way to the success of research into AI applications for healthcare.
I have over 10 years of experience in research and teaching in the field of Artificial intelligence, particularly Computer vision and Machine learning. I have led and been a key researcher of a number of research projects in the field, which resulted in more than 50 research papers published in international journals and conferences, 3 patents, and 10 graduate students supervised. Recently, recognizing the importance of AI applications in healthcare, I have set up a team of researchers and experts from AI and medical fields to work on a research project to develop an AI-assisted system for automated detection and classification of lesions in endoscopic images. We aimed at developing real-time Computer-aided diagnosis (CAD) systems as an “AI doctor” to aid endoscopists in reducing missed detection rates. We have obtained promising results in terms of the accuracy of the algorithms and the performance of the system. The success of this research will contribute to social impact on improving healthcare quality for people, especially in developing countries like Vietnam.
LinkedIn URL: https://www.linkedin.com/in/thuy-nguyen-51106027