Abhinav Gupta
agupta [at] cs.umd.edu

Graduate Research Assistant

Computer Vision Laboratory

University of Maryland- College Park

 

 

 



• About me
• Resume
• Research/Publications
• Selected Research Projects
• Downloads (Datasets/Code)
• Links

 News

  • Joining CMU as post-doc starting Aug 1 (with Alyosha Efros and Martial Hebert.)

  • Transactions on PAMI paper accepted 

                                                        

 

About Me

I am a fifth year PhD student in the Department of Computer Science at University of Maryland working with Prof. Larry Davis. I am also working in collaboration with Prof. Jianbo Shi at Grasp Lab, University of Pennsylvania.

My research interests include:

  • Computer Vision: Activity Analysis, Object Recognition, Image/Video Annotation, Motion and Tracking, Camera Networks, Shape and Appearance based Object Models.

 

  • Semantic Models: Storyline models, Language based contextual models, contextual models for interpreting human-object interactions.

 

  • Machine Learning: Learning from Weakly Labeled Data, Semi-Supervised Learning, Probabilistic Graphical Models for Recognition and Dimensionality reduction for human pose estimation.

 

  • Exploring synergy between language and vision for learning semantic models.

Resume/CV

 Resume (pdf)

Research/Publications*
 

Journal

  • Abhinav Gupta, Anurag Mittal and Larry S. Davis, Constraint Integration for Efficient Multiview Pose Estimation with Self-Occlusions, IEEE Transactions on Pattern Analysis and Machine Intelligence, 30(3), March 2008, pp. 493-506.  (PDF)
  • Abhinav Gupta, Aniruddha Kembhavi and Larry S. Davis, Observing Human-Object Interactions: Using Spatial and Functional Compatibility for Recognition. To Appear in IEEE Transactions on Pattern Analysis and Machine Intelligence (Special Issue on Probabilistic Graphical Models) (PDF)

Selected Conference Publications**

  • Abhinav Gupta, Praveen Srinivasan, Jianbo Shi and Larry S. Davis, Understanding Videos, Constructing Plots: Learning a Visually Grounded Storyline Model from Annotated Videos, In CVPR 2009. (Oral) (PDF)
  • Abhinav Gupta, Jianbo Shi and Larry S. Davis, A "Shape Aware" Model for semi-supervised Learning of Objects and its Context, In NIPS 2008 (Spotlight Poster) (PDF)
  • Abhinav Gupta and Larry S. Davis, Beyond nouns: Exploiting prepositions and comparative adjectives for learning visual classifiers, Accepted in ECCV 2008 (Oral) (PDF)
  • Abhinav Gupta, Trista Chen, Francine Chen, Don Kimber and Larry S. Davis, Context and Observation Driven Latent Variable Model for Human Pose Estimation, In CVPR 2008. (PDF)
  • Abhinav Gupta, Anurag Mittal and Larry S. Davis, COST*: An Approach for Camera Selection and Multi-Object Inference Ordering in Dynamic Scenes, In ICCV 2007 (PDF)
  • Abhinav Gupta and Larry S. Davis, Objects in Action:An Approach for Combining Action Understanding and Object Perception, In CVPR 2007 (PDF)

Others

  • Abhinav Gupta, Anurag Mittal and Larry S. Davis, Constraint Integration for Multiview Pose Estimation of Humans with Self-Occlusions, 3DPVT 2006 (PDF)
  • Abhinav Gupta, V. Shiv Naga Prasad and Larry S. Davis, Extracting Regions of Symmetry, 133-136, ICIP 2005, Genova, Italy (PDF)
  • Ashutosh Saxena, Abhinav Gupta and Amitabha Mukerjee, Non-Linear Dimensionality Reduction by Locally Linear Isomaps, ICONIP 2004, Published in Lecture Notes in Computer Science (PDF)
  • Abhinav Gupta and Phalguni Gupta, Watermarking of MPEG-4 Videos, ICBA 2004, Published in Lecture Notes in Computer Science, July 2004 (PDF)

Patents

  • Abhinav Gupta, Trista Chen, Francine Chen and Don Kimber. Observation-Driven Probability-Based Latent Models for Human Pose Estimation, Patent Filed, US Patent Application 11/955,270, 2007.

Downloads (Datasets)***

The following datasets can be for non-commercial research and educational purposes. Please cite the appropriate references when using the dataset in your research papers.

Please Register for the dataset first. The login and password will be displayed upon registration.

Multiple Person Tracking Datasets
This dataset has been used in M2Tracker, ICCV 2007, PAMI 2007 paper. It consist of videos consisting of 3, 4, 5 and 6 people in a scene. The dataset is available from 15cameras and calibration information is also provided.



 
Human-Object Interaction Dataset
This datasets consist of 10 subjects performing six activities. The dataset has been used in our CVPR 2007 paper.

 

 

Weakly Labeled Car Dataset - NIPS 2008 (Images from Label me)    Coming Soon !
This dataset consists of 45 training images for car and 48 test images. This dataset was used in our NIPS 2008 paper. We used the dataset in weakly labeled format (i.e positions of cars are not known in the training).

 

 

Action Recognition from Static Images    Coming Soon !
This dataset consists of 50 static images of different actions in different sports. List of actions are: batting (cricket), bowling (cricket), serve (tennis), forehand (tennis), serve (volleyball), shot (croquet). This dataset has been used in our submitted PAMI paper.

 
 

Synergistic Activities

 

Program Chair  

Workshop on Visual and Contextual Learning from Annotated Images and Videos (VCL'09)  (co-organized with Jianbo Shi and David Forsyth)  

 

Program Committee

Workshop on Visual Scene Understanding (with CVPR'09)  

Workshop on Use of Context in Vision Processing (with ICMI'09)

 

 

Links
 

Research
Computer Vision Home Page
Computer Vision Conferences
Group Page

Interesting
Restaurant Reviews and Food Recipes

Personal - My Friends
Swati Jarial
Anuj Rawat 
Srikanth Vishnubhotla
Aniruddha Kembhavi
Vlad Morariu
Pavan Kumar Turaga 

*Disclaimer: These documents are made available as a means to ensure timely dissemination of scholarly and technical work on a non-commercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each copyright holder. These works may not be reposted without the explicit permission of the copyright holder.

 

** Acceptance rates for Orals <5%, Spotlight ≈ 10-15% and Posters ≈ 20-25%

***The Downloads may not be complete as of yet. Please contact Abhinav Gupta for more information about the datasets.

Copyright © 2007 Abhinav Gupta. All Rights Reserved.