English Version 中文版

Prof. Dr. Xiaogang Cheng (成孝刚)

Now, I am an Associate Professor of computer vision and M. S. student supervisor at Nanjing University of Posts and Telecommunications(NUPT).


Research interests: 1)Vision-based non-invasive perception for human thermal comfort 2)Foggy and hazy visibility perception, Reflection removal.


Experience: 1) Study@Computer Vision Lab, ETH Zürich. Supervisor: Prof. Luc Van Gool, 2) Study@KTH Royal Institute of Technology, Sweden, 3) Study@Nanjing University and @Southeast University.


Contact: 1)E-mail: chengx@vision.ee.ethz.ch, xiacheng@kth.se, Chengxg@njupt.edu.cn, 2)Phone: +86 138 1337 2706


Welcome to join my academic research group!

[Google Scholar Profile]

Research interests

Main interests: Artificial Intelligence (AI) and Computer vision (CV)


Topic 1. Non-invasive thermal comfort perception for humanistic intelligent building

[1]  Project: Non-invasive thermal comfort perception based on deep learning.
[2]  Methodology: Machine learning (Deep Learning), computer vision.
[3]  Interdisciplinary subjects: Computer vision, machine learning and building physics.
[4]  Motivation: Energy saving, Human-centered indoor environment (buildings, vehicles).
[5]  Philosophical idea: Thermal comfort through perception.
[6]  Cooperation institutions: ETHZ Switzerland, KTH Sweden, UMU Sweden, LBNL USA, XAUAT China

Topic 2. Defogging and fogy visibility estimation for intelligent transportation

[1]  Project: Hazy visibility estimation and defogging based on deep learning and computer vision.
[2]  Methodology: Machine learning (Deep Learning), computer vision.
[3]  Interdisciplinary subjects: Computer vision, intelligent transportation and atmospheric sciences.
[4]  Motivation: Reduce traffic accidents, serve self-driving cars.
[5]  Cooperation institutions:  ETHZ Switzerland, KTH Sweden.

Main academic contributions

Build a new research sub-direction between computer vision and building physics.  For overcoming the drawbacks of current methods (invasive or semi-invasive, non-human-centered), we proposed vision-based non-invasive perception for human thermal comfort in Oct. 2016.

Construct algorithms. For overcoming the corresponding challenges, several algorithms were proposed. The challenges of human thermal comfort perception are 1) inter-individual difference, 2) intra-individual difference, and 2) subtle variation of skin texture.

Publications (Recent 3 years, first autor)

NIDL: A pilot study of contactless measurement of skin temperature for intelligent building((SCI, JCR Q1, cited: 2))
Xiaogang Cheng, Bin Yang, Anders Hedman, Thomas Olofsson, Haibo Li, Luc Van Gool
Energy & Buildings, 2019.
[ Paper ] [ Code ]
A contactless measuring method based on subtleness magnification and deep learning (NIDL) was designed to achieve a comfortable, energy efficient built environment.
A Contactless Measuring Method of Skin Temperature based on the Skin Sensitivity Index and Deep Learning(SCI, JCR Q4)
Xiaogang Cheng, Bin Yang, Kaige Tan, Erik Isaksson, Liren Li, Anders Hedman, Thomas Olofsson,Haibo Li
Appl. Sci. 2019
[ Paper ] [ Code ]
A contactless measuring method based on a skin sensitivity index and deep learning (NISDL) was proposed to measure real-time skin temperature.
A pilot study of online non-invasive measuring technology based on video magnification to determine skin temperature(SCI, JCR Q1, cited: 17)
Xiaogang Cheng, Bin Yang, Thomas Olofsson, Guoqing Liu, Haibo Li
Building and Environment, 2017.
[ Paper ] [ Code ]
Vision-based contactless perception method for human thermal comfort was proposed in the first time and subtleness magnification algorithm was adopted.
A variational approach to atmospheric visibility estimation in the weather of fog and haze(SCI, JCR Q2, cited: 4)
Xiaogang Cheng, Bin Yang, Guoqing Liu, Thomas Olofsson, Haibo Li
Sustainable Cities and Society, 2018
[ Paper ] [ Code ]
A variational framework to handle the nature of time-varying extinction coefficient and develop a novel algorithm of extracting the extinction coefficient through a piecewise functional fitting of observed luminance curves.
A Total Bounded Variation Approach to Low Visibility Estimation on Expressways(SCI, JCR Q3)
Xiaogang Cheng, Bin Yang, Guoqing Liu, Thomas Olofsson, Haibo Li
Sensors, 2018
[ Paper ] [ Code ]
A total bounded variation (TBV) approach to estimate low visibility (less than 300 m) is introduced.

Publications (Recent 3 years, second or corresponding author)

Non-Invasive Assessments of Thermal Discomfort in Real Time
Alan Meier, Xiaogang Cheng, William Dyer, Chris Graham, Thomas Olofsson, Bin Yang
COMFORT AT THE EXTREMES 2019.
[ Paper ] [Code]
A new method was introduced to assess a person's thermal discomfort based on close observation of human gestures.
Real-time and contactless measurements of thermal discomfort based on human poses for energy efficient control of buildings
Bin Yang, Xiaogang Cheng, Dengxin Dai, Thomas Olofsson, Haibo Li, Alan Meier
Building and Environment, 2019
[ Paper ] [ Code]
This paper examined a contactless method for evaluating a person's thermal sensation revealed by their poses.

Teaching award

Dec 2015     Excellent supervisor of NUPT
May 2015     3S Cup College Student competition in internet of things     Third prize (supervisor)
May 2016     3S Cup College Student competition in internet of things     Second prize (supervisor)

Graduates and employers

Master student: Kai Zhou(2014, @Media Tek), Miaomiao Tan(2014, @China mobile), Yun Cheng(2015, @Baidu), Hongjun Lv(2016, @Tongcheng Group), Dezhi Li(2016, @Huawei), Tao Wang(2016, @SpreadTrum), Lichang Zhang of KTH(2016, @Kunlun), Kaige Tan of KTH(2017, Ph.D.@KTH).

Bachelor student:  Zhi Li (2012, Ph.D@UB), Nan Huo(2015, Master@JHU), Yuchen Yang(2015, @UPenn), Chen Sun(2015, Master@Waterlo(UW)), Lin Zhu(2015, Master@NUPT), Peng Zheng(2015, Master@UniTrento).

Self-estimate

Outgoing, diligent, strong willed and sober personality, Good communication skill and teamwork spirit, Enjoying the challenges of academic study, Untalented but perseverant and indomitable.

Under construction.Last Updated on 18th Jan., 2020

Published with GitHub Pages