Personal Info

Wei Li is a Ph.D student at CUNY City College. He is advised by Prof. Zhigang Zhu. His research topic includes computer vision, deep learning, human compueter interaction and assistive techology for visually impaired people.

Google Scholar.

Github.

Email: BUAAWELLDON AT GMAIL DOT COM

Addr: NAC 8/210, Convent Ave 160, New York

Life

Will join Amazon AWS Computer Vision Research team, enjoy the days lefted at school...

One paper is accepted to FG 2017, see you in DC, May 29th.

Attend CVPR 2016 at Las Vegas, knew many new friends, also won money!!!

Will give a talk at Yahoo Research New York on Emotion recognition and Data Collection at Feb 11th.

Work as an intern for SRI International (2015 Dec --> 2016 Mar) on Assistive Driving projects.

A paper is accepted by ICMI 2015 , See you Seattle ^-^

REM Summer mentor(July,15 to August, 15)

    firefox
  • To help visually impaired people able to recognize others' emotions in their daily life, we developed a realtime app on android, check it out at google play store by scanning the bar code.
  • A web page is also developed, you can upload your image to test how it works at emotest.

Participate in the EmotiW challenge 2015

  • We are 4th/75 groups, cheers!

A paper published at MVA conference

Attend CVPR 2015 at Boston.

Projects

Facial evaluation for helping ASD(Autism Spectrum Disordered)

  • By using CNN and MKL based approach, extract feature from wild image and video, use the feature to analyze whether ASD people are reacting in a right pattern. The project is under construction.

Emotion Games

  • This is a game different with most of the games you saw. You control it just using your face. Match the icon shown up and you get points. try out the general version at here. or register to play a customized version at here.

Facial reading for Visually Impaired People

  • Using smart Phones as 'eyes' for visually impaired people to 'see' people around, knowing 'who' and 'how' are they looking? CNN based deep learning is employed to do the face recognition and expression recognition.

Deep Learning in expression recognition at IBM China Research Lab

  • Collecting more than 10k image data from online and using CAFFE train the expression Deep learning model. Comparing with the state of the art work based on SVM in both CK+ and our own dataset. Results show our approach have robust performance.

Publications

Papers

  • Wei Li, Farnaz Abtahi, Zhigang Zhu, Lijun Yin. EAC-Net: A Region-based Deep Enhancing and Cropping Approach for Facial Action Unit Detection. FG 2017, DC.pdf.
  • Wei Li, Farnaz Abtahi, Christina Tsangouri, Zhigang Zhu. Towards An "In-The-Wild" Emotion Dataset Using a Game-Based Framework. CVPRW 2016, Las Vegas.pdf. DATA: GaMo, CIFE.
  • Wei Li, Farnaz Abtahi, Zhigang Zhu. A deep feature based multi-kernel learning approach for video emotion recognition. 17th ACM International Conference on Multimodal Interaction (ICMI), Seattle, USA, 2015.pdf.
  • Farnaz Abtahi, Wei Li, Zhigang Zhu, Tony Ro. Multimodal Speaker Recognition using Deep Belief Networks, CVPRW 2015.
  • Li, Wei, et al. "A deep-learning approach to facial expression recognition with candid images." Machine Vision Applications (MVA), 2015 14th IAPR International Conference on. IEEE, 2015.pdf.
  • Li, Wei, et al. "Face Recognition by 3D Registration for the Visually Impaired Using a RGB-D Sensor." Computer Vision-ECCV 2014 Workshops. Springer International Publishing, 2014.
  • Li, Xudong, Wei Li, Hongzhi Jiang, and Huijie Zhao. "Automatic evaluation of machining allowance of precision castings based on plane features from 3D point cloud." Computers in Industry 64, no. 9 (2013): 1129-1137.
  • Wei LI, Xudong LI, Huijie ZHAO, Ying ZHANG. An Attitude Normalization Based Line Feature Point Cloud Extraction Approach. Journal of Beijing University of Aeronautics and Astronautics.
  • Li, Wei, et al. "Multiple View Point Cloud Registration Based on 3D Lines." Proceedings of the 2012 International Conference on Image Processing, Computer Vision Pattern Recognition, Las Vegas, USA. 2012.

Patents

  • Xudong LI, Huijie ZHAO, Wei LI, Hongzhi JIANG. A 3D Hough Transform based 3D Space Lines Recognition Method, NO. 201210210773.3, China Invention.
  • Xudong LI, Huijie ZHAO, Wei LI, Hongzhi JIANG. An Attitude Normalization Based 3D Feature Extraction Method From 3D Data. NO.201210209643.8, China Invention.
  • Huijie ZHAO, Xudong LI, Wei LI, Hongzhi JIANG. A 3D Data Based Workpiece Machining Allowance Estimation Method. NO. 201210210510.2, China Invention.
  • Xudong LI, Huijie ZHAO, Wei LI, Hongzhi JIANG. A 3D image Registration Approach Based On Space Lines Recognition. NO. 201210397752.7, China Invention.