Lam Huynh, Ph.D.

Moikka! I am the machine learning team lead at CubiCasa and a researcher at the Center for Machine Vision and Signal Analysis (CMVS), University of Oulu, Finland. Got my doctoral degree in Computer Science and Engineering from CMVS where I've co-advised by Prof. Janne Heikkilä, Prof. Jiří Matas, and Prof. Esa Rahtu working in the Infotech Oulu funded project.

Pique interest in large-scale 3D reconstruction, 3D scene understanding, 3D human pose, novel view synthesis, LLMs and LMMs. Love books, skiing, music, Jung's psychological types and effective learning.

Email
Github
Google Scholar
Curriculum vitae
LinkedIn
Stack Overflow
Research
Fast Neural Architecture Search for Lightweight Dense Prediction Networks
Lam Huynh, Esa Rahtu, Jiří Matas, Janne Heikkilä
arXiv, 2022
Lightweight Monocular Depth with a Novel Neural Architecture Search Method
Lam Huynh, Phong Nguyen-Ha, Jiří Matas, Esa Rahtu, Janne Heikkilä
WACV, 2022
Monocular Depth Estimation Primed by Salient Point Detection and Normalized
Hessian Loss

Lam Huynh, Matteo Pedone, Phong Nguyen-Ha, Jiří Matas, Esa Rahtu, Janne Heikkilä
3DV, 2021 (oral)
StressNAS: Affect State and Stress Detection Using Neural Architecture Search
Lam Huynh, Tri Nguyen, Thu Nguyen, Susanna Pirttikangas, Pekka Siirtola
ACM UbiComp-ISWC Adjunct, 2021
Boosting Monocular Depth Estimation with Lightweight 3D Point Fusion
Lam Huynh, Phong Nguyen-Ha, Jiří Matas, Esa Rahtu, Janne Heikkilä
ICCV, 2021
project page / arXiv / code
Guiding Monocular Depth Estimation Using Depth-Attention Volume
Lam Huynh, Phong Nguyen-Ha, Jiří Matas, Esa Rahtu, Janne Heikkilä
ECCV, 2020
project page / arXiv / code
Structure-from-motion using convolutional neural networks
Master thesis, funded by Cubicasa Oy and Business Finland
Published on Jultika, 2018
Supervisors: Prof. Janne Heikkilä, Dr. Markus Ylimäki

This work introduces a deep-learning-based structure-from-motion pipeline for the dense 3D scene reconstruction problem.

Projects
3D human pose estimation using deep neural networks
Visiting Researcher, IMD lab NAIST, Japan. July-October 2018.

Working on Human 3D Pose Estimation for Sprint (running) using learning-based approaches. This work is a collaboration with MSc. Yuma Ouchi, MSc. Taisei Watanabe, MSc. Ayaka Matsumoto and, Prof. Takafumi Taketomi.

Reading notes
Camera poses and coordinates for dummies
Documentation, Oulu, March 2019

This document is for everyone who still confuse with the relative camera poses and the relationship betweeen the camera and world coordinate. How to do the transformation from one camera coordinate to the other camera coordinate. There are many ways to do this wrong, therefore we should clearly understand these terms.

University pedagogy training
Documentation, Oulu, 30th March 2019

Learning diary of the basic of university pedagogy, from which I learned the fundamental of teaching and had lot of interesting discussions. Thanks to Dr. Sonja Lutovac for this amazing course.

The four Rs technique to become a deep learner
Documentation, Oulu, 05th April 2019

A practical scheme to remember and understand deeply what we have learned.

Teaching

Spring 2018:     Deep learning (Dr. Li Liu, Dr. Jie Chen)
Autumn 2019: Deep learning (Dr. Li Liu, Dr. Jie Chen)
Autumn 2020: Deep learning (Dr. Li Liu)
Autumn 2021: Deep learning (Dr. Li Liu)
I work as a Teaching Assisstant alongside MSc. Zhuo Su ('18-'21), Dr. Mohammad Tavakolian ('20), MSc. Yawen Cui ('20), MSc. Jiehua Zhang ('21), MSc. Huali Xu ('21), and MSc. Wuti Xiong ('21).

Academic Activities

Conference: ICME'21, WACV'22, ICME'22, WACV'23, AAAI'23, ICASSP'23, ICME'23, WACV'24, AAAI'24, ICASSP'24

Journal: IEEE Transactions on Pattern Analysis and Machine Intelligence, Image and Vision Computing, Neurocomputing, IET Computer Vision


The credit of this website template goes to Jon Barron. Thank you!
The curriculum vitae icon made from Icon Fonts is licensed by CC BY 3.0