An Tran
Tran Lam An, Trần Lâm Ẩn
PhD Candidate Computer Vision researcher
Email : an.tran [at]
Google Scholar, Github, LinkedIn, Google+.

The pinnacle of TECHNOLOGY is INNOVATION, that of SCIENCE is DISCOVERY, and more importantly that of EDUCATION is HUMANITY. Du Tran, June 2011.


Now, I am a Computer Vision researcher at AIQ, Singapore. I have finished my PhD study in ECE at NUS. My advisors are Cheong Loong-Fah and Qi Zhao. I did my bachelor of Computer Engineering in CSE at HCMUT.


  • 13-12-2017: I moved to AIQ as a Computer Vision researcher
  • 21-08-2017: One paper has been accepted in ICCVW 2017
  • 11-10-2015: One paper has been accepted in ICCV 2015
  • 11-10-2015: My homepage on github launched!


My general interests are in the fields of computer vision, machine learning and heterogeneous computing. Currently, my research focus is on action recognition, learning video representations and deep learning.

Two-stream Flow-guided Convolutional Attention Networks for Action Recognition
An Tran, and Cheong Loong-Fah
International Conference of Computer Vision Workshop (ICCVW), 2017
[ Arxiv ] [ Poster ] [ Supplementary Material ] [ BibTex ] [ Code ]
A new perspective of two-stream CNN with compensated optical flows.
Actionness-assisted Recognition of Actions
Ye Luo, Cheong Loong-Fah, and An Tran
International Conference of Computer Vision (ICCV), 2015
[ Paper ] [ BibTex ] [ Project Page ]

Challenges and Notebook Papers

  • NUS Submission to ActivityNet Challenge 2017: Evaluating frame-based and spatio-temporal CNN features for video classification
    An Tran and Cheong Loong-Fah
    ActivityNet Large Scale Activity Recognition Challenge, in conjuction with CVPR, 2017.
    [ PDF ]

Recent Talks


Graduate Assistant for

EE2024, Programming for Computer Interfaces ( Spring 2013 - Fall 2015 ).


On pleasure time, I usually read, watch and hopefully contribute to open source projects (e.g. OpenCV, Caffe).

Free Web Counter since Oct 2015