Gemma Roig

Ass. Prof. at  SUTD

Research Affiliate at MIT

Computational Vision  and Artificial Intelligence

Note: I am looking for students to work on a project related to Computational Human Vision and Deep Learning. If you are interested, please send me and email to gemmar@mit.edu.

Prospective PhD students:

-A*STAR I2R Thematic PhD programme

-SUTD PhD programme

 if you fulfill the prerequisites of a fellowship and are interested in working with me, please apply directly to the fellowships in the above links, indicating me as supervisor.

 

 

 

Bio

My research aim is to build computational models of human vision to understand its underlying principles, and to use those models to build applications of artificial intelligence.

 

I am currently an assistant professor at Singapore University of Technology and Design. I am also a research affiliate at MIT. Previously, I was a postdoctoral fellow at MIT in the Center for Brains Minds and Machines, with  Prof. Tomaso Poggio as my faculty host. I was also affiliated at the Laboratory for Computational and Statistical Learning, which is a collaborative agreement between the Istituto Italiano di Tecnologia and the Massachusetts Institute of Technology.  I pursued my doctoral degree in Computer Vision at ETH Zurich. Previously, I was a research assistant at the Computer Vision Lab at EPFL in Lausanne, at the Department of Media Technologies at Ramon Llull University in Barcelona, and at the Robotics Institute - Carnegie Mellon University in Pittsburgh.

Publications

Google scholar profile

 

  • Representation Similarity Analysis for Efficient Task Taxonomy and Transfer Learning.
  • K. Dwivedi, G. Roig.
  • Computer Vision and Pattern Recognition (CVPR), 2019. (accepted)
  •  
  • Joint Semantic-Instance Segmentation of 3D Point Clouds Using Multi-Set Label Conditional Random Fields.
  • Q.H. Pham, B.S. Hua, T. Nguyen, G. Roig, S.K. Yeung.
  • Computer Vision and Pattern Recognition (CVPR), 2019. (accepted)
  •  
  • Task-specific Vision Models Explain Task-specific Areas of Visual Cortex.
  • K. Dwivedi, G. Roig.

 

  • Navigational Affordance Cortical Responses Explained by Semantic Segmentation model.
  • K. Dwivedi, G. Roig.
  • In Proc. of European Conference on Computer Vision Workshops (ECCVW),  2018,
  • Brain-Driven Computer Vision (BDCV), 2018.
  • [pdf][bib]

 

  • DOPING: Generative Data Augmentation for Unsupervised Anomaly Detection with GAN.
  • S.K. Lim, Y. Loo, N.-T. Tran, N.-M. Cheung, G, Roig, Y. Elovici.
  • In Proc. of IEEE International Conference on Data Mining, 2018.
  • [pdf][bib]

 

  • Do Deep Neural Networks Suffer from Crowding?
  • A. Volokitin, G. Roig, T. Poggio.
  • In Proc. of Neural Information Processing Systems (NIPS), 2017.
  • [pdf][bib][code]

 

  • Eccentricity Dependent Deep Neural Networks: Modeling Invariance in Human Vision.
  • F. Chen, G. Roig, L. Isik, X. Boix and T. Poggio.
  • In AAAI Spring Symposium Series, Science of Intelligence, 2017.
  • [pdf][bib]

 

  • Is the Human Visual System Invariant to Translation and Scale?
  • Y. Han, G. Roig, G. Geiger and T. Poggio.
  • In AAAI Spring Symposium Series, Science of Intelligence, 2017.
  • [pdf][bib]

 

  • Herding Generalizes Diverse M-Best Solutions.

        E. Ozcan, G. Roig, O. Goksel, X. Boix.

        In arxiv:1611.04353, 2016.

        [pdf][bib]

 

  • Learning to Predict Sequences of Human Visual Fixations.
  • M. Jiang, X. Boix, G. Roig, J. Xu, L. Van Gool, and Q. Zhao.
  • In IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2016.
  • [pdf][bib]

 

  • Foveation-based Mechanisms Alleviate Adversarial Examples.
  • Y. Luo, X. Boix, G. Roig, T. Poggio, Q. Zhao.
  • CBMM memo 044, arXiv:1511.06292, 2015.
  • [pdf][bib]

 

  • SEEDS: Superpixels Extracted via Energy-Driven Sampling.

 

  • Saliency Prediction with Active Semantic Segmentation.
  • M. Jiang, X. Boix, J. Xu, G. Roig, L. Van Gool, Q. Zhao.
  • In Proc. of British Machine Vision Conference (BMVC), 2015.
  • [pdf][bib]

 

  • Self-Adaptable Templates for Feature Coding.
  • X. Boix*, G. Roig*, Salomon Diether, L. Van Gool.
  • In Proc. of Neural Information Processing Systems (NIPS), 2014.
  • (*equal contribution). [pdf][bib]

 

  • Random Binary Mappings for Kernel Learning and Efficient SVM.
  • G. Roig*, X. Boix*, L. Van Gool.
  • In arXiv:1307.5161, 2014.
  • (*equal contribution). [pdf][bib]
  •  
  • Comment on "Ensemble Projection for Semi-supervised Image Classification".
  • X. Boix, G. Roig, L. Van Gool.
  • In  arXiv:1408.6963, 2014.
  • [pdf][bib]

 

  • Active MAP Inference in CRFs for Efficient Semantic Segmentation.
  • G. Roig*, X. Boix*, R. de Nijs, S. Ramos, K. Kuehnlenz, L. Van Gool.
  • In Proc. of International Conferenence on Computer Vision (ICCV), 2013.
  • (Oral, *equal contribution). [pdf][bib]

 

  • Online Video SEEDS for Temporal Window Objectness.
  • M. Van den Bergh, G. Roig, X. Boix, S. Manen, L. Van Gool.
  • In Proc. of International Conferenence on Computer Vision (ICCV), 2013.
  •  [pdf][bib][code][project]

 

  • Sparse Quantization for Patch Description.
  • X. Boix, M. Gygli, G. Roig, L. Van Gool.
  • In Proc. of Computer Vision and Pattern Recognition (CVPR), 2013.
  • [pdf][bib]

 

  • Nested Sparse Quantization for Efficient Feature Coding.
  • X. Boix*, G. Roig*, L. Van Gool.
  • In Proc. of European Conferenence on Computer Vision (ECCV), 2012.
  • (*equal contribution). [pdf][bib]

 

  • SEEDS: Superpixels Extracted via Energy-Driven Sampling.
  • M. Van den Bergh, X. Boix, G. Roig, B. de Capitani, L. Van Gool.
  • In Proc. of European Conferenence on Computer Vision (ECCV), 2012.
  • [pdf][bib][OpenCV code][OpenCV Award][project]

 

  • On-line Semantic Perception Using Uncertainty.
  • R. de Nijs, S. Ramos, G. Roig, X. Boix, L. Van Gool, K. Kühnlenz.
  • In Proc. of International Conference on Intelligent Robots and Systems (IROS), 2012.
  • [pdf][bib]

 

  • Robotic ADaptation to Humans Adapting to Robots: Overview of the FP7 project RADHAR.
  • E. Demeester, E. Vander Poorten, A. Hüntemann, J. De Schutter, B. Lau, M. Kuderer, W. Burgard, A. Fossati, G. Roig, X. Boix, M. Ristin, L. Van Gool, et al.
  • In Proc. of 1st International Conference on Systems and Computer Science, 2012.
  • [pdf][bib][project]

 

  • Conditional Random Fields for Multi-Camera Object Detection.
  • G. Roig*, X. Boix*, H. Ben Shitrit, P. Fua.
  • In Proc. of International Conferenence on Computer Vision (ICCV), 2011.
  • (*equal contribution). [pdf][bib][project]

 

  • Hierarchical CRF with Product Label Spaces for Parts-based Models.
  • G. Roig*, X. Boix*, F. De la Torre, J. Serrat, C. Vilella.
  • In Proc. of Face and Gesture (FG) 2011.
  • (Oral, *equal contribution). [pdf][bib]

 

  • Optimal Feature Selection for Subspace Image Matching.
  • G. Roig, X. Boix, F. De la Torre.
  • In Proc. of International Conference on Computer Vision Workshops (ICCVW), 2009.
  • [pdf][bib]

 

 

 

       INTERNATIONAL WORKSHOPS AND CONFERENCE ABSTRACTS

 

  •  
  • Eccentricity Dependent Neural Network with Recurrent Attention for Scale, Translation and Clutter Invariance.
  • J. Zhang, Y. Han, T. Poggio, G. Roig.
  • Vision Sciences Society Conference, 2019.
  • (Abstract and talk).

 

  • Explaining Scene-selective Visual Area Using Task-specific and Category Specific DNN Units.
  • K. Dwivedi, M. Bonner, G. Roig.
  • Vision Sciences Society Conference, 2019.
  • (Abstract + poster).

 

  • Properties of invariant object recognition in human one-shot learning suggests a hierarchical architecture different from deep convolutional neural networks.
  • Y. Han, G. Roig, G. Geiger, T. Poggio.
  • Vision Sciences Society, 2019.
  • (Abstract + poster)

 

  • Importance of object selection in Relational Reasoning tasks.
  • K. Dwivedi, G. Roig.
  • Neural Information Processing Systems Workshops, 2018
  • Relation Representation Learning (R2L 2018).

 

  • Plug and Play DNN Modules for Multi-domain Learning.
  • K. Dwivedi, G. Roig.
  • European Conference on Computer Vision Workshops (ECCVW), 2018,
  • Interactive and Adaptive Learning in an Open World (IAL 2018).

 

  • Do Deep Neural Networks Suffer from Crowding?
  • G. Roig, A. Volokitin, T. Poggio.
  • Vision Science Society, 2018.
  • ([Abstract] + talk).
  •  
  • Blacklisted speaker Identification Using Tripel Neural Networks.
  • K.W. Cheuk, B.T. Balamurali, G. Roig, D. Herremans.
  • In IEEE International Conference on Data Mining (ICDM), 2018
  • 1st Multi-target speaker detection and identification Challenge Evaluation, 2018.
  • ([Abstract]).
  •  
  • Eccentricity Dependent Deep Neural Networks for Modeling Human Vision.

        G. Roig, F. Chen, X. Boix and T. Poggio.

        Vision Sciences Society, 2017.

       ([Abstract] + poster).

 

  • On the Human Visual System Invariance to Translation and Scale.
  • Y. Han, G. Roig, G. Gaiger and T. Poggio.
  • Vision Sciences Society, 2017.

        ([Abstract] + poster).

 

  • Online Video SEEDS for Temporal Window Objectness.
  • M. Van den Bergh, G. Roig, X. Boix, S. Manen, L. Van Gool.
  • In European Conferenence on Computer Vision (ECCVW), 2013.
  • (Talk + poster).

 

  • Ensemble Kernel Learning.
  • G. Roig*, X. Boix*, C. Leistner, L. Van Gool.
  • In International Conferenence on Computer Vision Workshops (ICCVW), 2011.
  • (Abstract + poster).

 

 

Contact

Gemma Roig, PhD

 

gemmar@mit.edu

 

Singapore University of Technology and Design

Information Systems Technology and Design Pillar

8 Somapah Road,

Building 1, level 7

Singapore 487372



 

Home

 GEMMA ROIG

 

GEMMA ROIG