MPI-INF Logo
Homepage

Research Interests / Bio

UPDATE! I started as a Research Scientist at Google (Zurich) in February 2022. My new website will be onlie soon. Students -- do not hesitate to reach out to me in case you are seeking collaboration on GAN based rendering and Human Centric Vision.

My research goal is to assist in creating digital media for AR/VR and editing applications that are indistinguishable from reality. This includes, but is not limited to, algorithms to create digital and virtual humans with limited data (up to a single image) that are photorealistic and fully controllable (pose, viewpoints, shape, appearance, etc.), estimation and generation of human pose and gesture, virtual try on, neural rendering, and 3D representations.

Bio. I am was a postdoctoral researcher in the Visual Computing and AI department (previously GVV) of Prof. Christian Theobalt at Max-Planck Institute for Informatics, where I work on photorealistic rendering of humans and human-centric vision. I obtained my PhD under Prof. Didier Stricker at German Research Center for Artificial Intelligence (DFKI) Kaiserslautern, where I designed 3D shape representations for general machine learning applications. I got my bachelor and masters degrees in the subject of Computer Science and Engineering from the Indian Institute of Technology Kharagpur, India (IIT Kharagpur).

During my undergraduate/postgraduate studies, I interned as a Google Summer of Code Student for Point Cloud Library (2012) and RoboComp (2015), and as a summer student at University of Tuebingen and University of Western Ontario. My C++ library (created from scratch) OpenDetection was selected as a Google Summer of Code Participating Organization for two consecutive years (2016-2017).

Publications



    Style and Pose Control for Image Synthesis of Humans from a Single Monocular View.
    K. Sarkar, V. Golyanik, L. Liu and C. Theobalt.
    arXiv.org, 2021.
    [paper] [project page]

    HumanGAN: A Generative Model of Humans Images
    K. Sarkar, L. Liu, V. Golyanik, and C. Theobalt.
    arXiv.org, 2021.
    [paper] [project page]
    Neural Actor: Neural Free-view Synthesis of Human Actors with Pose Control.
    L. Liu, M. Habermann, V. Rudnev, K. Sarkar, J. Gu and C. Theobalt.
    arXiv.org, 2021.
    [paper] [project page]

    EgoRenderer: Rendering Human Avatars from Egocentric Camera Images
    T. Hu, K. Sarkar, L. Liu, M Zwicker and C. Theobalt.
    International Conference on Computer Vision (ICCV), 2021
    [paper] [project page]

    Estimating Egocentric 3D Human Pose in Global Space
    J. Wang, L. Liu, W. Xu, K. Sarkar and C. Theobalt.
    International Conference on Computer Vision (ICCV), 2021
    [paper] [project page]

    Pose-Guided Human Animation from a Single Image in the Wild.
    J. S. Yoon, L. Liu, V. Golyanik, K. Sarkar, H. S. Park, and C. Theobalt.
    Computer Vision and Pattern Recognition (CVPR), 2021
    [paper] [video] [project page]


    Simple domain adaptation for CAD based object recognition.
    K. Sarkar and D. Stricker.
    International Conference on Pattern Recognition Applications and Methods, 2019.
    [paper]


    Structured low-rank matrix factorization for point-cloud denoising.
    K. Sarkar, F. Bernard, K. Varanasi, C. Theobal, and D. Stricker.
    International Conference on 3D Vision (3DV), 2018
    [paper] [supplement]


    Learning 3d shapes as multi-layered height-maps using 2d convolutional networks.
    K. Sarkar, B. Hampiholi, K. Varanasi, and D. Stricker.
    European Conference on Computer Vision (ECCV), 2018.
    [paper] [supplement]


    Denoising of Point-clouds Based on Structured Dictionary Learning.
    K. Sarkar, F. Bernard, K. Varanasi, C. Theobal, and D. Stricker.
    Symposium on Geometry Processing, Eurographics Association (SGP), 2018. (invited poster)
    [paper]


    3D Shape Processing by Convolutional Denoising Autoencoders on Local Patches.
    K. Sarkar, K. Varanasi, and D. Stricker.
    Winter Conference on Applications of Computer Vision (WACV), 2018
    [paper]


    Learning quadrangulated patches for 3d shape parameterization and completion.
    K. Sarkar, K. Varanasi, and D. Stricker.
    International Conference on 3D Vision (3DV), 2017
    [paper]


    Trained 3d models for CNN based object recognition.
    K. Sarkar, K. Varanasi, and D. Stricker.
    International Conference on Computer Vision Theory and Applications (VISAPP), 2017
    [paper]


    Feature-augmented trained models for 6dof object recognition and camera calibration.
    K. Sarkar, A. Pagani, and D. Stricker.
    International Conference on Computer Vision Theory and Applications (VISAPP), 2016
    [paper]

Teaching

  • Summer 2019
    2D Image Processing (2DIP 2019), TU Kaiserslautern - Teaching Assistant and Lecturer (60% of the lectures).
  • Summer 2018
    2D Image Processing (2DIP 2019), TU Kaiserslautern - Teaching Assistant and Lecturer (30% of the lectures).

Recent Positions

  • Since August 2019
    Postdoc at Max-Planck-Institut for Informatics
  • 2015 - July 2019
    Researcher at German Center for Artificial Intelligence (DFKI)
  • 2012 - 2014
    Software Engineer at Paypal India

Education

  • 2015 - July 2019
    PhD student at TU Kaiserslautern. Status - Thesis submitted; currently waiting for reviews.
  • 2007 - 2012
    Bachelor(H) and Master of Technology in Computer Science and Engineering
    Indian Institute of Technology, Kharagpur (IIT Kharagpur)