Xê-mi-na khoa học
22/02/2022: TS. Nguyễn Hoàng Hà (ĐH KH&CN, Hà Nội ), Automatic morphometric landmark detection on insect wing images
The study of biological forms increasingly utilizes the geometric morphometrics approach as it allows quantitative evaluation of morphology. Shapes could be digitized as a set of landmarks from specimen images. The annotation of landmarks on insect wings is manually done by domain experts; however, it is time- consuming and error-prone. In this talk, I will present our work on automating this task by using both the traditional image processing method and the deep learning approach. On the one hand, our customized CNN network and some object detected deep learning models can give high accurate landmark recognition. On the other hand, we propose a key- point-matching-based process that is lightweight to run, requires small sample datasets and is realized into an open- source tool. Our hand-crafted feature framework is accurate with reference to human work.
Speaker: TS. Nguyễn Hoàng Hà, ĐH KH&CN, Hà Nội
Time: 15:30, Tuesday, February 22, 2022
Venue: Webinar
Dr. NGUYEN Hoang Ha is currently a lecturer and researcher at the Information and Communication Technology department, University of Science and Technology of Hanoi. He received his bachelor (honor program) and master’s degrees from the University of Engineering and Technology, VNU. His Ph.D. thesis on realistic road model reconstruction from GIS data was successfully defended at Aix-Marseille University, France in 2016. His research interests include Computer Graphics and Vision, Machine Learning, and AI. He has published some papers in 3D modeling and Video Processing fields. He has also collaborated with the industrial sector to carry out R&D projects on software engineering and camera surveillance. He has received prestigious awards such as Young Face of the Year of VNU, Microsoft Most Valuable Professional.