Burlingame, California, United States
• Built a detector that detects the 3D positions of facial landmarks on a 3D scan of a real person.
• Using R3DS Wrap and the 3D facial landmark detector, individually developed a fully automatic
pipeline registering temporal sequences of 3D scans of a person to a mesh template.
• Independently designed a fully automatic pipeline that augment temporal sequences of scans using
Open3D, putting glasses on 3D scans of a person as if the person is wearing the glasses. The
augmented sequence is also temporally smoothed for reduction of jittering.
• Built a quantitative evaluation method for the augmentation pipeline, which evaluates temporal
jittering of the glasses and verifies the validity of the wearing styles.