Abir Das

Welcome to my webpage

Hi, I'm Abir Das. I am an Assistant Professor in the Computer Science and Engineering Department at IIT Kharagpur. Before joining IIT Kharagpur I was briefly an Assistant professor at IIIT Delhi. I was a postdoctoral researcher at Boston University under Prof. Kate Saenko before that. I completed my PhD in Electrical Engineering from the  University of California at Riverside (UCR). My research advisor was Dr. Amit Kumar Roy-Chowdhury in the 'Video Computing Group' led by him. Before joinging UCR, I had done my Bachelor of Electrical Engineering from Jadavpur University, Calcutta, India in the year 2007.

Please visit the 'Research' page of this website to go over the research problems that I have explored in the past and am exploring recently.

PhD Positions
I am looking for highly motivated and talented PhD students interested to do do research on computer vision and machine learning. My interests lie in reinforcement learning, explainable AI and bias in Deep Learning models for Computer Vision. Interested students may look at my publications page and the research group page to get a glimpse of the kind of problems I am interested in.

Recent News:

  • Paper - "Semi-Supervised Action Recognition with Temporal Contrastive Learning" accepted in CVPR 2021.
  • Organizing CVPR 2021 Workshop on Dynamic Neural Networks Meets Computer Vision (DNetCV).
  • 2020:
  • Paper - "Mitigating Dataset Imbalance via Joint Generation and Classification" accepted in ECCV 2020 workshop on Imbalance Problems in Computer Vision (IPCV).
  • Paper - "Reinforcement Learning Based Load Balancing for Hybrid LiFi WiFi Networks" accepted in IEEE Access.
  • Paper - "Revisiting Few-shot Activity Detection with Class Similarity Control" accepted in IEEE CVPR 2020 workshop on Visual Learning with Limited Labels.
  • 2019:
  • Paper - "Two-Stream Region Convolutional 3D Network for Temporal Activity Detection" accepted in IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI).
  • 2018:
  • One paper accepted in BMVC 2018 as oral (4.3% acceptance rate).
  • Named as an outstanding reviewer of CVPR 2018 .
  • 2017:
  • Our work 'R-C3D: Region Convolutional 3D Network for Temporal Activity Detection' won the award for most innovative solution in ActivityNet challenge in CVPR 2017.
  • Two papers accepted in ICCV 2017.
  • One paper accepted in CVPR 2017.
  • One paper accepted in CVIU.
  • 2016:
  • Our work 'Multimodal Video Description' won third place in MSR-VTT challenge in ACM-MM 2016.
  • One paper accepted in ECCV 2016.
  • One paper accepted in PAMI.