Tong Xu

Tong Xu

Master student of Computer Science

University of Southern California

Biography

I am a second-year Master student at University of Southern California. I spent wonderful several months as visiting researcher in Human-to-Everything (H2X) Lab at the beautiful Boston University, fortunately supervised by Prof. Eshed Ohn-Bar. I also worked closely with my friend Zanming Huang and Dr. Jimuyang Zhang. My research interests are object recognition, visual odometry and related applications in autonomous driving and robotics manipulation.

Previously, I received my bachelor’s degree in Network Engineering at Nanjing University of Information Science & Technology with an honor degree in June, 2021. I am lucky to work under the supervision of Prof. Xiaorui Zhang at NUIST.

Interests
  • Autonomous Vehicles
  • Visual Odometry
  • Object Recognition
  • Cognitive Robotics
Education

  • M.S. in Computer Science, 2021

    University of Southern California

  • B.S. in Network Engineering, 2017

    Nanjing University of Information Science & Technology

News

  • 06/2022: We finished the project OpenGuide - A Scalable Human-Like Guidance System for Travelers with Visual Impairment!
  • 05/2022: I joined Human-to-Everything (H2X) Lab as visiting researcher, supervised by Prof. Eshed Ohn-Bar.
  • 08/2021: I started my master study at University of Southern California.
  • 06/2021: I graduated from Nanjing University of Information Science & Technology!

Experience

 
 
 
 
 
H2X Lab
Visiting Researcher
H2X Lab
May 2022 – Present Boston

Responsibilities include:

  • Publishing papers
  • Analysing Visual Odometry tasks
  • Helping OpenGuide Project

Selected Publications

Quickly discover relevant content by filtering publications.
Xiaorui Zhang, Tong Xu, Wei Sun, Aiguo Song (2021). Multiple source domain adaptation in micro-expression recognition. Journal of Ambient Intelligence and Humanized Computing.

PDF Cite Source Document DOI

Xiaorui Zhang, Wenfang Zhang, Wei Sun, Tong Xu, Sunil Kumar (2020). A Robust Watermarking Scheme Based on ROI and IWT for Remote Consultation of COVID-19. Computers, Materials & Continua.

PDF Cite Source Document DOI

Tong Xu (2019). Single-view and Multi-view Methods in Marker-less 3D Human Motion Capture. Journal of Physics - Conference Series.

PDF Cite Source Document DOI

Selected Projects

*
All Deep Learning Visual Odometry Data Structure
DeepVO
Visual odometry tasks with FlowNet pretrained structure and LSTM models.
Source code
DeepVO
Cone Detection
Reproduced Faster R-CNN model by combining Region Proposal Network with Fast R-CNN. Compared the performance of Faster R-CNN model with YOLOv3 model on cone annotated dataset.
Source code Slides Document
Cone Detection
OpenGuide
A Scalable Human-Like Guidance System for Travelers with Visual Impairment.
Source code Demo
OpenGuide
Sequence Alignment
Class Project - Sequence Alignment using the basic Dynamic Programming approach and a space-efficient version of the Hirschberg Algorithm.
Source code
Sequence Alignment
SRCNN
Accomplished image super-resolution reconstruct tasks by SRCNN model.
Source code
SRCNN

Contact