Data augmentation is an effective technique for improving the accuracy and reducing over-fitting of modern image classifiers. However, current data augmentation implementations are manually designed. There are several papers been publishing to tackle this problem such as, AutoAugment (Cubuk et al, CVPR 2019), Population Based Augmentation (Ho et al., ICML 2019) and Fast AutoAugment (Lim et al., NeurIPS 2019), by automatically find the optimal augmentation policies. In this seminar, the aforementioned papers are presented. Furthermore, we customized the Fast AutoAugment to apply the method into pathology prediction from X-ray images.
Dr.Nguyễn is one the key members of the Applied Scientist team at VinBrain. He previously got his Bachelor of Engineering degree at HCMC University of Technology. Then Dr.Nguyễn spent 7 years in Korea to get his MSc and Ph.D. degree at SeoulTech and KAIST, respectively. His research interests are robot vision, 3D reconstruction, and image enhancement.
WeWork coworking space - Floor 24th, 11 Đoàn Văn Bơ, District 4, Hồ Chí Minh city
Other Events You May Like
Conference Room (Floor 7th), Vinmec International Hospital, Vinhomes Times City, 458 Minh Khai, Hai Bà Trưng district, Hà Nội.
Almaz Center, Vinhomes Riverside, Long Bien, Hanoi