Welcome to my website. I’m a first-year Ph.D. student of Electrical Engineering and Computer Science at UC Merced PANS LAB, advised by Prof. Shijia Pan. I obtained my B.S. degree and M.A. degree in Statistics from Nanjing University and Columbia University in 2021 and 2022, respectively. My research interests include machine learning, deep learning, and their applications in various fields.
In daily life, I love singing songs and playing the guitar. Please visit my channel if you are interested.
sdu14 AT ucmerced dot edu
IOT Classification Challenge: Activity Detection
Understanding human activities is an essential part in building a smart IOT environment. This project aims to use vision-based methods to classify human activities, different interactions with a fridge, in an IOT scenario. We use detectron2 to extract a skeleton representation of human gestures, which is simplified to the coordinates of 17 keypoints within a frame. The skeleton data extracted from a video is a time series. So we use the Transformer-based deep learning method to model the temporal information.