Ishan Nigam

I am a Masters student in the Robotics Institute at Carnegie Mellon University. Previously, I worked as a Research Assistant at IIIT Delhi India after graduating from the University of Delhi.

At CMU, I pursue research in Computer Vision advised by Dr. Deva Ramanan. My general interests lie in understanding how high level Vision tasks can be modeled for imperfect data distributions.

Email  /  CV  /  Google Scholar  /  Github

Ensemble Knowledge Transfer for Semantic Segmentation, IEEE WACV 2018 [PDF, Github]
Ishan Nigam, Chen Huang, Deva Ramanan
Network ensembles trained on complementary visual domains outperform network enembles trained on singular domains.
Learning Universal Embeddings with Attributes [PDF]
Chen Huang, Ishan Nigam, Deva Ramanan
Multi-task framework to learn universal embeddings supervised by triplet similarity; trained robustly with an adversarial net.
Exploring the Structure of Residual Neural Networks [PDF]
Dozen-layer wide residual networks outperform deep residual networks of equivalent capacity for object classification.
Temporal Scene Completion with GANs [PDF]
Conditional image synthesis and view synthesis with appearance flow for hole filling in video sequences.
  1. Ensemble Knowledge Transfer for Semantic Segmentation, IEEE WACV 2018 [PDF]
    I. Nigam, C. Huang, D. Ramanan
  2. [Submitted] Effect of Cataract Surgery on Iris Pattern Recognition, Nature Eye.
    I. Nigam, R. Keshari, M. Vatsa, R. Singh, K. Bowyer
  3. Ocular Biometrics: A Survey of Fusion and Modalities, Elsevier Information Fusion, Vol. 26, Nov 2015. [IF: 5.7] [PDF]
    I. Nigam, M. Vatsa, R. Singh
  4. Ophthalmic Disorder Menagerie and Iris Recognition, Springer's Handbook of Iris Recognition, Vol. 2, 2015. [PDF]
    I. Nigam, M. Vatsa, R. Singh
  5. Revisiting HEp-2 Cell Classification, IEEE Access, Vol. 3, 3102-3113, 2015. [IF: 3.2] [PDF]
    I. Nigam, S. Agrawal, M. Vatsa, R. Singh
  6. A Leap Password based Verification System, 7th IEEE BTAS 2015, Best Poster. [PDF]
    A. Chahar, S. Yadav, I. Nigam, M. Vatsa, R. Singh
  7. Leap Signature Recognition using HOOF and HOT features, 21st IEEE ICIP 2014 [PDF]
    I. Nigam, M. Vatsa, R. Singh