Mar 2021 - Present
Face Recognition
Developing and deploying face recognition system on edge device.
Mainly responsible for
1. Developing a framework to conduct face data augmentation (e.g. face's pose, eyeglasses, smile, etc.) from existing dataset.
2. Redesigning and Improving face detection model in the application field of wearing mask (RMSE of landmark with mask: 3.514 -> 3.45; wo mask: 2.926 -> 2.606).
3. Improving accuracy of face recognition model from 96.61 to 97.32.
4.Deploying face recognition system on IOS.
Fall Detection
Developing and deploying fall detection systems in the field of self-driving car and medical care.
Mainly responsible for
1. Fine-tuning pre-trained model in different application fields.
2. Deploying AI model on Jetson nano.