I am a first-year master student in Robotics Institute at Carnegie Mellon University.

Previously, I received Bachelor of Science degree in Computer Science and Mathematics from Hong Kong University of Science and Technology with First Class Honors, where I worked with Prof. Chi-Keung Tang and Prof. Yu-Wing Tai. I also spent several months in Bytedance AI Lab working on self-supervised video recognition with Dr. Qi She and in MEGVII video detection group wroking on object detection.

My research interests are computer vision and deep learning. Particularly, I am highly interested in solving real-world perception problems with scarce or unlabeled data, which covers the topics of self-supvervised learning, few-shot learning, object/video classification detection. I am also interested in autonomous driving.

Email  /  Google Scholar  /  Github  /  LinkedIn

profile photo
Carnegie Mellon University
Master of Science in Computer Vision | Aug. 2021 - Dec. 2022(Expected)
Hong Kong University of Science and Technology
Bachelor of Science in Computer Science and Mathematics | Sep. 2016 - Jun. 2020 | GPA: 3.96/4.30
ETH Zurich
Undergraduate Exchange in Computer Science | Sep. 2018 - Jun. 2019

(* indicates equal contribution)

Inter-intra Variant Dual Representations for Self-supervised Video Recognition
Lin Zhang, Qi She, Zhengyang Shen, Changhu Wang
BMVC, 2021
project page / arXiv

Dual representations to learn inter-intra variance of videos through self-supervision

MT-ORL: Multi-Task Occlusion Relationship Learning
Panhe Feng, Qi She, Lei Zhu, JiaXin Li, Lin ZHANG, Zijian Feng, Changhu Wang, Chunpeng Li, Xuejing Kang, Anlong Ming
ICCV, 2021

Porposed a unified framework and novel loss functions for occlusion reasoning, including both boundary detection and occlusion relationship prediction.

One-Shot Object Detection without Fine-Tuning
Xiang Li*, Lin Zhang*, Yau Pun Chen, Yu-Wing Tai, Chi-Keung Tang
arxiv, 2020
project page / arXiv

A two stage detection pipepline and architecture for one shot object detection with metric learning style

NeRF Speedup at Test Time via Super-resolution
Lin Zhang, Qifeng Chen

Train a super-resolution model on low-high resolution pairs generated by NeRF to accelerate NeRF generation at test time

Video Interpolation and Extrapolation in Long Intervals
Lin Zhang, Yu-Wing Tai, Chi-Keung Tang
slides / videos

Joint video interpolation and extrapolation on Cityscape dataset within long time intervals

Duplicate Question Detection
Lin Zhang, Dit-Yan Yeung
tech report

Detect duplicate questions on Quora dataset using RNN sequential models and autoencoder-generated augmented data

Analytics and Recommendation for User Location Data
Lin Zhang, Zijian Huang Gary Shueng Han CHAN
tech report

Designed and built an indoor localization system with Android devices, beacons, database and web development.

Work Experiences
ByteDance AI Lab
Feb. 2021 - Jun. 2021 | Dr. Qi She

Topic: self-supervised human action recognition
Jan. 2020 - Apr. 2020

Topic: object detection
HKUST Robotics Team
Nov. 2016 - Jul. 2017 | Prof. Kam-Tim Woo

Topic: camera-guided balance car design