Carlos Esteves

Table of Contents

Home

Bio

I am a Research Scientist at Google Research, NYC.

I completed my Ph.D. in Computer and Information Science at the GRASP Laboratory, University of Pennsylvania, under the supervision of Kostas Daniilidis.

Prior to that I received the B.S. in computer engineering and M.S. in electronic and computer engineering degrees from the Aeronautics Institute of Technology (ITA) in Brazil.

photo.png

News

  • Jul 19, 2021: Our new paper, Implicit-PDF: Non-Parametric Representation of Probability Distributions on the Rotation Manifold has been accepted to ICML and will be presented this week! See the project page for all details!
  • Sep 26, 2020: Our paper, Spin-Weighted Spherical CNNs has been accepted to NeurIPS! Camera-ready here, code coming soon!
  • Sep 13, 2020: I was invited to speak at the Workshop on Equivariance and Data Augmentation and the recordings are now available. Check out my talk on Spin-Weighted Spherical CNNs here!
  • Apr 13, 2020: I wrote a manuscript about the theory behind Group CNNs, "Theoretical Aspects of Group Equivariant Neural Networks". Check it out here.
  • Sep 09, 2019: The extended version of our paper Learning SO(3) Equivariant Representations with Spherical CNNs will appear on IJCV "Best of ECCV 2018" special edition. New version is already available here.
  • Jul 22, 2019: Our paper, Equivariant Multi-View Networks, has been accepted as oral to ICCV 2019! Acceptance rate was 4.6%. Paper here.
  • Apr 23, 2019: Our paper, Cross-domain 3D Equivariant Image Embeddings, has been accepted to ICML! Camera-ready available here.
  • Jul 05, 2018: Our paper, Learning SO(3) Equivariant Representations with Spherical CNNs, has been accepted as an oral presentation to ECCV 2018! Acceptance rate was 2.4%. Camera-ready version here.
  • Jan 31, 2018: Our Polar Transformer Networks paper has been accepted at ICLR 2018, with top 3% review scores!

Research

My research is on computer vision and machine learning. I'm interested in:

  • equivariant representations,
  • harmonic analysis,
  • 3D vision,
  • pose estimation,
  • wide baseline image matching.

Selected publications

Implicit-PDF: Non-Parametric Representation of Probability Distributions on the Rotation Manifold
K. Murphy, C. Esteves, V. Jampani, S. Ramalingam, A. Makadia
International Conference on Machine Learning, ICML 2021.
[ pdf | code | project page ]

Spin-Weighted Spherical CNNs
C. Esteves, A. Makadia, K. Daniilidis.
Conference on Neural Information Processing Systems, NeurIPS 2020.
[ pdf | code | talk ]

Equivariant Multi-View Networks
C. Esteves, Y. Xu, C. Allen-Blanchette, K. Daniilidis.
International Conference on Computer Vision, ICCV 2019 (oral; acceptance rate 4.6%).
[ pdf | poster | talk | code ]

Cross-domain 3D Equivariant Image Embeddings
C. Esteves, A. Sud, Z. Luo, K. Daniilidis, A. Makadia.
International Conference on Machine Learning, ICML 2019.
[ pdf | poster ]

Learning SO(3) Equivariant Representations with Spherical CNNs
C. Esteves, C. Allen-Blanchette, A. Makadia, K. Daniilidis.
European Conference on Computer Vision, ECCV 2018 (oral; acceptance rate 2.4%).
International Journal of Computer Vision, "Best of ECCV 2018" special edition, (IJCV 2019).
[ pdf | poster | talk | code | IJCV extended version ]

Labeling Panoramas with Spherical Hourglass Networks
C. Esteves, K. Daniilidis, A. Makadia.
360° Perception and Interaction Workshop at ECCV 2018.
[ pdf | poster ]

Polar Transformer Networks
C. Esteves, C. Allen-Blanchette, X. Zhou, K. Daniilidis.
International Conference on Learning Representations, ICLR 2018.
[ pdf | poster | code | OpenReview ]

An Analysis of SVD for Deep Rotation Estimation
J. Levinson, C. Esteves, K. Chen, N. Snavely, A. Kanazawa, A. Rostamizadeh, A. Makadia
Conference on Neural Information Processing Systems, NeurIPS 2020.
[ pdf ]

Fast Multi-Image Matching via Density-Based Clustering
R. Tron, X. Zhou, C. Esteves, K. Daniilidis.
International Conference on Computer Vision, ICCV 2017.
[ pdf ]

Manuscripts

Learning Equivariant Representations
Carlos Esteves
Ph.D. Thesis
[ pdf ]

Theoretical Aspects of Group Equivariant Neural Networks
Carlos Esteves
[ pdf ]

Service

  • CVPR reviewer (IEEE/CVF Conference on Computer Vision and Pattern Recognition)
    • CVPR 2019 Outstanding Reviewer
    • CVPR 2021 Outstanding Reviewer
  • ICML reviewer (International Conference on Machine Learning)
    • ICML 2021 Expert Reviewer
  • NeurIPS reviewer (Conference on Neural Information Processing Systems)
  • ICCV reviewer (IEEE International Conference on Computer Vision)
  • ECCV reviewer (European Conference on Computer Vison)
  • TPAMI reviewer (IEEE Transactions on Pattern Analysis and Machine Intelligence)
  • AAAI reviewer (AAAI Conference on Artificial Intelligence)
  • GMDL reviewer (Geometry Meets Deep Learning Workshop)
  • 360PI reviewer (360o Perception and Interaction Workshop)

Teaching

Math for Machine Learning, MCIT-515. Teaching assistant, Summer/2020.
Machine Perception, CIS-580. Teaching assistant, Spring/2018, Spring/2019.
Advanced Robotics, MEAM-620. Teaching assistant, Spring/2017.
Robotics: Planning and Perception, CIS-390. Teaching assistant, Fall/2016.

External

Author: Carlos Esteves

Created: 2021-07-19 Mon 12:26

Validate