Shiyi Lan

I am an undergraduate student [at] School of Computer Science, Fudan University, China. I was mentored by Prof. Wei Zhang and
Prof. Xiangyang Xue when I was doing research about Object Detection and Segmentation. I am a research intern in Megvii(Face++), mentored by Mr. Yuning Jiang and supervised by Dr. Jian Sun, focusing on research about Object Detection and Segmentation. Now, I am doing research about Video Classification and Representation, mentored by Prof. Fei Sha and Prof. Yugang Jiang. My research interests lie in the field of Machine Learning, Computer Vision, Object Detection and Segmentation, 3D Reconstruction, Video Representation and Classification.


FastMask: Segment Multi-scale Object Candidates in One Shot

Hexiang Hu*, Shiyi Lan*, Yuning Jiang, Zhimin Cao, Fei Sha (*Equal contribution)
Appeared as a spotlight paper at CVPR 2017 in Honolulu, Hawaii


Currently, I am visiting the lab of FeiSha [at] USC and designing and experimenting models of video classification and representation.

FeiSha [at] USC

In 2016, I joined Megvii(Face++), a leading Chinese AI start-up, as a research intern, mentored by Mr. Yuning Jiang, and supervised by Dr. Jian Sun.


In 2015, I participated in ACM/ICPC and got No.17 (Silver First).


In 2014, I became an undergraduate student majoring in Computer Science in Fudan University, China.

Fudan Univeristy

Fun Projects

A camera application on iOS that can apply neural style filter to photos. A SqueezeNet pretrained on ImageNet and MXNet ported to iOS are used in this project. I solved many compatibility issues in the project and my pull request to these issues for MXNet is accepted by MXNet Official Development Group.

Neural Style iPhone Camera

A Youtube-like video site for students to watch, search, upload and share videos. AngularJS and Django are used in the project including a uploader supporting resuming from breakpoint, a danmaku(rolling comments) system, a video searcher and many good-looking pages.

ChannelV on Fudan Student Website

A web wishing wall where students write, read and comment their wishes. ReactJS, Tornado, MongoDB are used in this project. The server I implemented processes each request asynchronously so that it can handle over 500 requests per second.

Wishing Wall on Fudan Student Website