ModiFace, Toronto, Canada | May 2023 – Dec 2023

Machine Learning Intern

Conducted literature review, reproduced existing methods, and made research proposals.

Proposed a novel self-supervised facial landmark estimation method and implemented it with PyTorch.

Carried out extensive experiments and demonstrated our proposed method outperforming existing SOTA by 20%~44% on landmark matching and 9%~15% on landmark detection.

Composed a conference paper which is accepted to CVPR 2024 as the first author.