Georgios Pavlakos
pavlakos@cs.utexas.edu

I am an Assistant Professor in the Department of Computer Science at the University of Texas at Austin. Before that, I was a Postdoctoral Researcher at UC Berkeley, advised by Angjoo Kanazawa and Jitendra Malik. I completed my PhD in Computer Science at the University of Pennsylvania with my advisor, Kostas Daniilidis. I did my undergraduate studies at the National Technical University of Athens, where I worked with Petros Maragos. During my PhD, I spent time at the Max Planck Institute in Tübingen, working with Michael Black.

Email  /  CV  /  Google Scholar  /  GitHub

• New papers accepted to CVPR 2024! HaMeR, BUDDI, GART and MultiPhys!
• I started as an Assistant Professor of Computer Science at UT Austin in January 2024!
• 4D Humans accepted to ICCV 2023!
• New papers accepted to CVPR 2023! SLAHMR, LART and Animals3D!
• I will be an Area Chair for CVPR 2024, ICCV 2023, BMVC 2023 and 3DV 2024!
• New paper accepted to ECCV 2022!
• New papers accepted to CVPR 2022! PHALP and Multishot!
• Interview at Computer Vision News - Best of ICCV selection!
• I was recognized as Outstanding Reviewer for ICCV 2021 and 3DV 2021!
• I received the Morris and Dorothy Rubinoff Award for the Best Computer Science Dissertation at UPenn!
• I started as a Postdoctoral Researcher at UC Berkeley with Angjoo Kanazawa and Jitendra Malik!

Scaling up data and models for hand mesh recovery from images and video.

A 3D generative model of two people in close social interaction.

3D Gaussian Splatting for non-rigid articulated subject capturing and rendering from monocular videos.

Using a physics simulator to recover physically plausible multi-person 3D motion.

A fully "transformerized" deisgn for Human Mesh Recovery achieves improved precision and remarkable robustness for 3D human reconstruction and tracking!

Reasoning about the global motion of humans in the world by decoupling human and camera motion.

Using 3D human reconstruction and tracking to recognize atomic actions in video.

Training models for 3D animal recovery with minimal annotations using large-scale collections of web images.

Recovering 3D humans, cameras and static structure in TV shows.

Predicting pose, appearance and location of people in 3D for monocular tracking.

Using information from multiple shots to improve reconstruction of humans in edited media.

Performing monocular tracking of people by lifting them to 3D and then using 3D representations of their appearance, pose and location.

Casting human mesh recovery as a regression from an image to a distribution of 3D poses, and showing the benefits on downstream tasks.

Reactive navigation in unexplored environments cluttered with familiar or unknown objects.

Regression-based expressive capture of 3D humans from a single RGB image.

End-to-end reconstruction of multiple people using two novel geometric losses that encourage coherent 3D estimates.

Reactive human following with semantic targets in unexplored environments cluttered with familiar or unknown objects.

Leveraging texture consistency to train networks for human mesh estimation.

Improving model-based human pose and shape regression with automatic in-the-loop fitting

Expressive Body Capture: 3D Hands, Face and Body from a Single Image
Georgios Pavlakos*, Vasileios Choutas*, Nima Ghorbani, Timo Bolkart, Ahmed A. A. Osman, Dimitrios Tzionas, Michael J. Black
Computer Vision and Pattern Recognition (CVPR), 2019   (Oral Presentation)
project page / supplementary / video / code / bibtex

Expressive capture of bodies, hands and faces from a single RGB image.

Estimating 3D human pose and shape using Graph Convolutional Networks.

Ordinal Depth Supervision for 3D Human Pose Estimation
Georgios Pavlakos, Xiaowei Zhou, Kostas Daniilidis
Computer Vision and Pattern Recognition (CVPR), 2018   (Oral Presentation)
project page / supplementary / video / code / data / bibtex

Incorporating ordinal depth supervision in the training of end-to-end ConvNets for 3D human pose estimation.

Learning to Estimate 3D Human Pose and Shape from a Single Color Image
Georgios Pavlakos, Luyang Zhu, Xiaowei Zhou, Kostas Daniilidis
Computer Vision and Pattern Recognition (CVPR), 2018  
Invited to 3D HUMANS, CVPR Workshop, 2018   (Best Poster Award)
project page / supplementary / bibtex

A direct, end-to-end approach for the reconstruction of 3D human pose and shape from single images.

Human Motion Capture Using a Drone
Xiaowei Zhou, Sikang Liu, Georgios Pavlakos, Vijay Kumar, Kostas Daniilidis
International Conference on Robotics and Automation (ICRA), 2018
video / data / bibtex

Estimating the 3D human pose from a monocular video captured by a drone.

MonoCap: Monocular Human Motion Capture using a CNN Coupled with a Geometric Prior
Xiaowei Zhou, Menglong Zhu, Georgios Pavlakos, Spyridon Leonardos, Konstantinos G. Derpanis, Kostas Daniilidis
Pattern Analysis and Machine Intelligence (PAMI), 2018
code / bibtex

Estimating the 3D pose of a human from a monocular video using a ConvNet to localize 2D keypoints and an EM optimization scheme to recover 3D pose over time.

Coarse-to-Fine Volumetric Prediction for Single-Image 3D Human Pose
Georgios Pavlakos, Xiaowei Zhou, Konstantinos G. Derpanis, Kostas Daniilidis
Computer Vision and Pattern Recognition (CVPR), 2017   (Spotlight Presentation)
project page / supplementary / video / training code / demo code / bibtex

End-to-end learning for 3D human pose using a volumetric representation and casting the problem as 3D keypoint localization in a discretized 3D space.

Harvesting Multiple Views for Marker-less 3D Human Pose Annotations
Georgios Pavlakos, Xiaowei Zhou, Konstantinos G. Derpanis, Kostas Daniilidis
Computer Vision and Pattern Recognition (CVPR), 2017   (Spotlight Presentation)
project page / supplementary / video / code / bibtex

Estimating 3D human pose from multiple views and leveraging the estimates as automatic 3D annotations for human pose estimation tasks.

6-DoF Object Pose from Semantic Keypoints
Georgios Pavlakos, Xiaowei Zhou, Aaron Chan, Konstantinos G. Derpanis, Kostas Daniilidis
International Conference on Robotics and Automation (ICRA), 2017
project page / journal version / code / video / bibtex

Estimating the 6-DoF pose of an object from a single image using semantic keypoints and a deformable shape model.

Reconstruction of 3D Pose for Surfaces of Revolution from Range Data
Georgios Pavlakos, Kostas Daniilidis
International Conference on 3D Vision (3DV), 2015
bibtex

On Shape Recognition and Language
Petros Maragos, Vassilis Pitsikalis, Athanasios Katsamanis, Georgios Pavlakos, Stavros Theodorakis
Perspectives in Shape Analysis, 2016, edited by M. Breuss, A. Bruckstein, P. Maragos and S. Wuhrer
bibtex

Kinect-based multimodal gesture recognition using a two-pass fusion scheme
Georgios Pavlakos, Stavros Theodorakis, Vassilis Pitsikalis, Athanasios Katsamanis, Petros Maragos
International Conference on Image Processing (ICIP), 2014
bibtex

Towards an intelligent robotic walker for assisted living using multimodal sensorial data
Georgia Chalvatzaki, Georgios Pavlakos, Kevis Maninis, Xanthi S. Papageorgiou, Vassilis Pitsikalis, Costas S. Tzafestas, Petros Maragos
International Conference on Mobile Communication and Healthcare (Mobihealth), 2014
bibtex

Advances in intelligent mobility assistance robot integrating multimodal sensory processing
Xanthi S. Papageorgiou, Costas S. Tzafestas, Petros Maragos, Georgios Pavlakos, Georgia Chalvatzaki, George Moustris, Iasonas Kokkinos, Angelika Peer, Bartlomiej Stanczyk, Evita-Stavroula Fotinea, Eleni Efthimiou
International Conference on Universal Access in Human-Computer Interaction (HCII), 2014
bibtex

This guy has an awesome website