Staff Machine Learning Engineer at Dropbox

me@neerajkumar.org

Research Overview

I work at the intersection of computer vision and machine learning, developing techniques that leverage large collections of real-world images for a variety of applications. I am particularly interested in the use of intermediate representations such as parts and attributes, both for improving performance on classical vision tasks such as search and recognition, as well as for creating novel applications such as automatic face replacement and exploration of image collections. Much of my work is directed toward human faces, although I am also very interested in the growing area of fine-grained visual categorization, which seeks to distinguish categories within a particular domain. One such project is Leafsnap, an electronic field guide that lets users identify plant species by photographing their leaves. The freely-available apps have received over a million downloads.


neeraj headshot

News

September 09, 2014: I've joined Dropbox!
I'm thrilled to announce that after 3 wonderful years at UW working with Steve Seitz, I've now joined the Photos team at Dropbox, along with my PhD advisor Peter Belhumeur, his current student Tom Berg, and David Kriegman from UCSD!
July 11, 2014: Leafsnap Dataset released
The leafsnap dataset has been released.
May 09, 2014: Photo Recall paper accepted to Web-scale Vision and Social Media (VSM) workshop at CVPR 2014
Our paper Photo Recall: Using the Internet to Label Your Photos has been accepted to the 2nd Workshop on Web-scale Vision and Social Media (VSM) which will be held this June in conjunction with CVPR 2014.
July 25, 2013: LFPW dataset available again from my website
The Labeled Face Parts in the Wild (LFPW) dataset is temporarily available from my webpage, as the KBVT site on which it was previously hosted has been undergoing renovations.

Curriculum Vitae

CV (pdf)
Citations to my work

Talks (see all)

CSE 590V: Computer vision seminar (Fall 2011)

Selected Projects (see all)

Photo Recall: Using the Internet to Label Your Photos

Fall 2012 - Ongoing
Photo Recall: Using the Internet to Label Your Photos

Searching through personal photos without any manual labeling

Leafsnap: An Electronic Field Guide

Summer 2009 - Fall 2012
Leafsnap: An Electronic Field Guide

Automatically identifying plant species using photos of leaves taken in the field by mobile devices

Multi-Attribute Spaces: Calibration for Attribute Fusion and Similarity Search

Fall 2011 - Summer 2012
Multi-Attribute Spaces: Calibration for Attribute Fusion and Similarity Search

Calibrating the outputs of multiple attribute classifiers into normalized multi-attribute spaces for fusion and similarity search

Localizing Parts of Faces Using a Consensus of Exemplars

Fall 2009 - Summer 2011
Localizing Parts of Faces Using a Consensus of Exemplars

Robustly localizing face parts using local detectors and non-parametric, exemplar-based global models

Attribute and Simile Classifiers for Face Verification

Spring 2009 - Summer 2011
Attribute and Simile Classifiers for Face Verification

Identifying faces using describable visual attributes of the face, automatically detected using attribute and simile classifiers

FaceTracer: A Search Engine for Large Collections of Images with Faces

Spring 2008 - Summer 2011
FaceTracer: A Search Engine for Large Collections of Images with Faces

Searching face images using automatically-trained classifiers of descriptions of face and image attributes

Selected Software (see all)

Attribute Calibration Software

Summer 2012
Attribute Calibration Software

C/C++/Python code to calibrate attribute classifier outputs using the w-score formulation, based on the Extreme Value Theory. For non-commercial uses only.

Leafsnap

Spring 2011 - Fall 2012
Leafsnap

The Leafsnap homepage, with links to download the iPhone and iPad apps, as well as browse the list of species and their images.

MugHunt Online Face Search Engine

Spring 2010 - Fall 2012
MugHunt Online Face Search Engine

An online face search engine built using this research. This search engine is a joint collaboration with Securics, Inc..

Databases

Leafsnap Dataset

Spring 2011 - Fall 2012
Leafsnap Dataset

30,866 leaf images of 185 species along with automatically-generated segmentations

Labeled Face Parts in the Wild (LFPW)

Summer 2011
Labeled Face Parts in the Wild (LFPW)

1,432 Images with 29 points labeled on each image

PubFig: Public Figures Face Database

Fall 2009
PubFig: Public Figures Face Database

60,000 face images with several face verification benchmarks & 65 automatically computed attribute labels for 42,000 images

FaceTracer Database

Fall 2008
FaceTracer Database

15,000 face images with detailed pose and fiducial information & 5,000 manual attribute labels

Selected Publications (see all)

  1. "Photo Recall: Using the Internet to Label Your Photos,"
    Neeraj Kumar, Steven Seitz,
    2nd Workshop on Web-scale Vision and Social Media (VSM) at CVPR 2014,
    June 2014.
    [pdf] [bibtex]
    @InProceedings{photorecall_vsm2014,
    author = {Neeraj Kumar and Steven Seitz},
    title = {Photo Recall: Using the Internet to Label Your Photos},
    booktitle = {2nd Workshop on Web-scale Vision and Social Media (VSM) at CVPR 2014},
    month = {June},
    year = {2014}
    }
  2. "Photo Recall: Using the Internet to Label Your Photos," (extended abstract)
    Neeraj Kumar, Steven Seitz,
    Proceedings of the 23rd International Conference on World Wide Web companion,
    April 2014.
    [pdf] [bibtex] [poster]
    @InProceedings{photorecall_www2014,
    author = {Neeraj Kumar and Steven Seitz},
    title = {Photo Recall: Using the Internet to Label Your Photos},
    booktitle = {The 23rd International Conference on World Wide Web companion},
    month = {April},
    year = {2014}
    }
  3. "Localizing Parts of Faces Using a Consensus of Exemplars,"
    Peter N. Belhumeur, David W. Jacobs, David J. Kriegman, Neeraj Kumar,
    IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI),
    vol. 35, no. 12, pp. 2930--2940, December 2013.
    [pdf] [bibtex] [slides (ppt)] [poster]
    @InProceedings{fiducials_pami2013,
    author = {Peter N. Belhumeur and David W. Jacobs and David J. Kriegman and Neeraj Kumar},
    title = {Localizing Parts of Faces Using a Consensus of Exemplars},
    booktitle = {IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI)},
    volume={35},
    number={12},
    pages={2930-2940},
    month = {December},
    year = {2013}
    }
  4. "How reliable are your visual attributes?,"
    Walter Scheirer, Neeraj Kumar, Vijay N. Iyer, Peter N. Belhumeur, Terrance E. Boult,
    Proceedings of SPIE Volume 8712: Defense, Security, and Sensing,
    April 2013.
    [pdf] [bibtex]
    @InProceedings{reliability_spie2013,
    author = {Walter Scheirer and Neeraj Kumar and Vijay N. Iyer and Peter N. Belhumeur and Terrance E. Boult},
    title = {How reliable are your visual attributes?},
    booktitle = {Proceedings of SPIE Volume 8712: Defense, Security, and Sensing},
    volume={8712},
    month = {April},
    year = {2013}
    }
  5. "Leafsnap: A Computer Vision System for Automatic Plant Species Identification," (oral presentation)
    Neeraj Kumar, Peter N. Belhumeur, Arijit Biswas, David W. Jacobs, W. John Kress, Ida Lopez, João V. B. Soares,
    Proceedings of the 12th European Conference on Computer Vision (ECCV),
    October 2012.
    [pdf] [bibtex] [slides (ppt)]
    @InProceedings{leafsnap_eccv2012,
    author = {Neeraj Kumar and Peter N. Belhumeur and Arijit Biswas and David W. Jacobs and W. John Kress and Ida Lopez and João V. B. Soares},
    title = {Leafsnap: A Computer Vision System for Automatic Plant Species Identification},
    booktitle = {The 12th European Conference on Computer Vision (ECCV)},
    month = {October},
    year = {2012}
    }
  6. "Multi-Attribute Spaces: Calibration for Attribute Fusion and Similarity Search,"
    Walter Scheirer, Neeraj Kumar, Peter N. Belhumeur, Terrance E. Boult,
    Proceedings of the 25th IEEE Conference on Computer Vision and Pattern Recognition (CVPR),
    June 2012.
    [pdf] [bibtex] [poster]
    @InProceedings{multiattrs_cvpr2012,
    author = {Walter Scheirer and Neeraj Kumar and Peter N. Belhumeur and Terrance E. Boult},
    title = {Multi-Attribute Spaces: Calibration for Attribute Fusion and Similarity Search},
    booktitle = {The 25th IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
    month = {June},
    year = {2012}
    }
  7. "Fusing with Context: a Bayesian Approach to Combining Descriptive Attributes," (oral presentation)
    Walter Scheirer, Neeraj Kumar, Karl Ricanek, Terrance E. Boult, Peter N. Belhumeur,
    Proceedings of the IEEE International Joint Conference on Biometrics (IJCB),
    October 2011.
    [pdf] [bibtex]
    @InProceedings{fusion_ijcb2011,
    author = {Walter Scheirer and Neeraj Kumar and Karl Ricanek and Terrance E. Boult and Peter N. Belhumeur},
    title = {Fusing with Context: a Bayesian Approach to Combining Descriptive Attributes},
    booktitle = {The IEEE International Joint Conference on Biometrics (IJCB)},
    month = {October},
    year = {2011}
    }
  8. "Two faces are better than one: Face recognition in group photographs,"
    Ohil K. Manyam, Neeraj Kumar, Peter N. Belhumeur, David J. Kriegman,
    Proceedings of the IEEE International Joint Conference on Biometrics (IJCB),
    October 2011.
    [pdf] [bibtex] [poster]
    @InProceedings{twofaces_ijcb2011,
    author = {Ohil K. Manyam and Neeraj Kumar and Peter N. Belhumeur and David J. Kriegman},
    title = {Two faces are better than one: Face recognition in group photographs},
    booktitle = {The IEEE International Joint Conference on Biometrics (IJCB)},
    month = {October},
    year = {2011}
    }
  9. "Describable Visual Attributes for Face Verification and Image Search,"
    Neeraj Kumar, Alexander C. Berg, Peter N. Belhumeur, Shree K. Nayar,
    IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI),
    vol. 33, no. 10, pp. 1962--1977, October 2011.
    [pdf] [bibtex] [slides (ppt)] [video]
    @InProceedings{facever_pami2011,
    author = {Neeraj Kumar and Alexander C. Berg and Peter N. Belhumeur and Shree K. Nayar},
    title = {Describable Visual Attributes for Face Verification and Image Search},
    booktitle = {IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI)},
    volume={33},
    number={10},
    pages={1962-1977},
    month = {October},
    year = {2011}
    }
  10. "Describable Visual Attributes for Face Images," (PhD Thesis)
    Neeraj Kumar,
    Technical Report CUCS-035-11, Department of Computer Science, Columbia University,
    August 2011.
    [pdf] [bibtex] [slides (ppt)]
    @InProceedings{nkthesis,
    author = {Neeraj Kumar},
    title = {Describable Visual Attributes for Face Images},
    booktitle = {Technical Report CUCS-035-11, Department of Computer Science, Columbia University},
    month = {August},
    year = {2011}
    }
  11. "Localizing Parts of Faces Using a Consensus of Exemplars,"
    Peter N. Belhumeur, David W. Jacobs, David J. Kriegman, Neeraj Kumar,
    Proceedings of the 24th IEEE Conference on Computer Vision and Pattern Recognition (CVPR),
    June 2011.
    [pdf] [bibtex] [slides (ppt)] [poster]
    @InProceedings{fiducials_cvpr2011,
    author = {Peter N. Belhumeur and David W. Jacobs and David J. Kriegman and Neeraj Kumar},
    title = {Localizing Parts of Faces Using a Consensus of Exemplars},
    booktitle = {The 24th IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
    month = {June},
    year = {2011}
    }
  12. "Attribute and Simile Classifiers for Face Verification," (oral presentation)
    Neeraj Kumar, Alexander C. Berg, Peter N. Belhumeur, Shree K. Nayar,
    Proceedings of the 12th IEEE International Conference on Computer Vision (ICCV),
    October 2009.
    [pdf] [bibtex] [slides (ppt)]
    @InProceedings{facever_iccv2009,
    author = {Neeraj Kumar and Alexander C. Berg and Peter N. Belhumeur and Shree K. Nayar},
    title = {Attribute and Simile Classifiers for Face Verification},
    booktitle = {The 12th IEEE International Conference on Computer Vision (ICCV)},
    month = {October},
    year = {2009}
    }
  13. "FaceTracer: A Search Engine for Large Collections of Images with Faces,"
    Neeraj Kumar, Peter N. Belhumeur, Shree K. Nayar,
    Proceedings of the 10th European Conference on Computer Vision (ECCV),
    pp. 340--353, October 2008.
    [pdf] [bibtex] [slides (ppt)] [poster]
    @InProceedings{facesearch_eccv2008,
    author = {Neeraj Kumar and Peter N. Belhumeur and Shree K. Nayar},
    title = {FaceTracer: A Search Engine for Large Collections of Images with Faces},
    booktitle = {The 10th European Conference on Computer Vision (ECCV)},
    month = {October},
    year = {2008}
    }
  14. "What is a Good Nearest Neighbors Algorithm for Finding Similar Patches in Images?,"
    Neeraj Kumar, Li Zhang, Shree K. Nayar,
    Proceedings of the 10th European Conference on Computer Vision (ECCV),
    pp. 364--378, October 2008.
    [pdf] [bibtex] [poster]
    @InProceedings{nn_eccv2008,
    author = {Neeraj Kumar and Li Zhang and Shree K. Nayar},
    title = {What is a Good Nearest Neighbors Algorithm for Finding Similar Patches in Images?},
    booktitle = {The 10th European Conference on Computer Vision (ECCV)},
    month = {October},
    year = {2008}
    }
  15. "Face Swapping: Automatically Replacing Faces in Photographs," (oral presentation)
    Dmitri Bitouk, Neeraj Kumar, Samreen Dhillon, Peter N. Belhumeur, Shree K. Nayar,
    ACM Transactions on Graphics (Proceedings of SIGGRAPH),
    vol. 27, no. 3, August 2008.
    [pdf] [bibtex] [slides (ppt)] [slides (ppt+videos)]
    @InProceedings{facereplace_siggraph2008,
    author = {Dmitri Bitouk and Neeraj Kumar and Samreen Dhillon and Peter N. Belhumeur and Shree K. Nayar},
    title = {Face Swapping: Automatically Replacing Faces in Photographs},
    booktitle = {ACM Transactions on Graphics (Proceedings of SIGGRAPH)},
    volume={27},
    number={3},
    month = {August},
    year = {2008}
    }