Denis Rozumny
Note: my name can also be spelled Denys Rozumnyi.
I am a Research Scientist at Meta Reality Labs in Peter Kontschieder 's team.
I finished my PhD at the Computer Vision and Geometry Group, Department of Computer Science, ETH Zürich under supervision of Prof. Marc Pollefeys . I also worked closely with Prof. Martin Oswald , Prof. Vittorio Ferrari , and Prof. Jiri Matas .
Previously, I was a research intern at Google Research with Prof. Vittorio Ferrari and at Meta Reality Labs .
Before that, I finished my MSc and BSc degrees from CTU in Prague , Center for Machine Perception under supervision of Prof. Jiri Matas .
My research interests are in 3D reconstruction of objects and scenes, video understanding, realistic rendering, object detection, tracking, 6D pose estimation. In particular, I focus on test-time optimization methods to solve those problems.
Email /
CV /
Google Scholar /
Twitter /
GitHub /
LinkedIn
News
Nov 2024 : I've joined Meta Reality Labs as a Research Scientist.
Oct 2024 : My Meta internship work has been accepted to WACV'24.
Jul 2024 : One paper accepted to ECCV'24: Master thesis student who I supervised.
Jun 2024 : Won 1st place in the Structured Semantic 3D Reconstruction challenge at CVPR'24 workshop USM3D .
Mar 2024 : Finished my research internship at Meta.
Sep 2023 : My internship work at Google is accepted to NeurIPS'23.
July 2023 : Two papers accepted at ICCV'23.
Research
Your browser does not support the video tag.
Retrieval Robust to Object Motion Blur
Rong Zou ,
Marc Pollefeys ,
Denys Rozumnyi ,
ECCV , 2024
paper /
GitHub
We propose a method for object retrieval in images that are affected by motion blur.
Your browser does not support the video tag.
Single-Image Deblurring, Trajectory and Shape Recovery of Fast Moving Objects With Denoising Diffusion Probabilistic Models
Radim Spetlik ,
Denys Rozumnyi ,
Jiri Matas ,
WACV , 2024
paper /
GitHub
A diffusion-based model that deblurs and recovers shape of fast moving objects from a single image without a known background (for the first time).
Estimating Generic 3D Room Structures from 2D Annotations
Denys Rozumnyi ,
Stefan Popov ,
Kevis-Kokitsi Maninis ,
Matthias Nießner ,
Vittorio Ferrari
NeurIPS , 2023
arXiv /
reviews /
presentation /
poster /
poster source /
GitHub
We propose a novel method to produce generic 3D room layouts just from 2D segmentation masks, with which we annotate and publicly release 2246 3D room layouts on the RealEstate10k dataset.
Human from Blur: Human Pose Tracking from Blurry Images
Yiming Zhao ,
Denys Rozumnyi ,
Jie Song ,
Otmar Hilliges ,
Marc Pollefeys ,
Martin R. Oswald
ICCV , 2023
project page /
arXiv /
video /
poster
We estimate 3D human poses from substantially blurred images, e.g. extension of Shape from Blur to SMPL human body model.
Your browser does not support the video tag.
Tracking by 3D Model Estimation of Unknown Objects in Videos
Denys Rozumnyi ,
Jiri Matas ,
Marc Pollefeys ,
Vittorio Ferrari ,
Martin R. Oswald
ICCV , 2023
arXiv /
video /
poster /
poster source /
GitHub
We propose to guide and improve 2D tracking with an explicit object representation, namely the textured 3D shape and 6DoF pose in each video frame.
Finding Geometric Models by Clustering in the Consensus Space
Daniel Barath ,
Denys Rozumnyi ,
Ivan Eichhardt ,
Levente Hajder ,
Jiri Matas
CVPR , 2023
arXiv /
GitHub
A new algorithm for finding an unknown number of geometric models.
Motion-from-Blur: 3D Shape and Motion Estimation of Motion-blurred Objects in Videos
Denys Rozumnyi ,
Martin R. Oswald ,
Vittorio Ferrari ,
Marc Pollefeys
CVPR , 2022
arXiv /
video /
poster /
poster source /
GitHub
Extension of Shape from Blur to multiple frames with more complex trajectories and exposure time modeling.
Shape from Blur: Recovering Textured 3D Shape and Motion of Fast Moving Objects
Denys Rozumnyi ,
Martin R. Oswald ,
Vittorio Ferrari ,
Marc Pollefeys
NeurIPS , 2021
arXiv /
reviews /
presentation /
video /
poster /
poster source (keynote) /
GitHub (110 stars)
The first method to estimate textured 3D shape and sub-frame 6D motion of fast moving objects from a single frame.
Your browser does not support the video tag.
FMODetect: Robust Detection of Fast Moving Objects
Denys Rozumnyi ,
Jiri Matas ,
Filip Sroubek ,
Marc Pollefeys ,
Martin R. Oswald
ICCV , 2021
arXiv /
slides /
poster /
poster source /
GitHub
The first deep-learning based approach for fast moving object detection.
DeFMO: Deblurring and Shape Recovery of Fast Moving Objects
Denys Rozumnyi ,
Martin R. Oswald ,
Vittorio Ferrari ,
Jiri Matas ,
Marc Pollefeys
CVPR , 2021
project page /
arXiv /
video /
poster /
poster source /
GitHub (165 stars)
We propose DeFMO that given a single image with its estimated background outputs the object's appearance and position in a series of sub-frames as if captured by a high-speed camera (i.e. temporal super-resolution). This is the first deep-learning based approach for FMO deblurring.
Your browser does not support the video tag.
Tracking by Deblatting
Denys Rozumnyi ,
Jan Kotera ,
Filip Sroubek ,
Jiri Matas
IJCV , 2021
project page /
Springer /
GitHub
Summarization and extension of our GCPR'19 (TbD-NC) and ICCVW'19 (TbD) papers.
Sub-frame Appearance and 6D Pose Estimation of Fast Moving Objects
Denys Rozumnyi ,
Jan Kotera ,
Filip Sroubek ,
Jiri Matas
CVPR , 2020
project page /
arXiv /
video /
GitHub
We extend TbD pipeline to track fast moving objects in full 6 DoF, estimating simultaneously their 3D motion trajectory, 3D pose and object appearance changes with a time step that is a fraction of the video frame exposure time.
Your browser does not support the video tag.
Non-Causal Tracking by Deblatting
Denys Rozumnyi ,
Jan Kotera ,
Filip Sroubek ,
Jiri Matas
GCPR , 2019   (Oral Presentation, Best Paper Honorable Mention)
project page /
arXiv /
GitHub
We apply post-processing with dynamic programming and curve fitting to obtain more accurate object trajectories.
Your browser does not support the video tag.
Intra-frame Object Tracking by Deblatting
Jan Kotera ,
Denys Rozumnyi ,
Filip Sroubek ,
Jiri Matas
ICCVW , 2019
project page /
arXiv /
GitHub
We propose a novel approach called Tracking by Deblatting to track fast moving objects.
Learned Semantic Multi-Sensor Depth Map Fusion
Denys Rozumnyi ,
Ian Cherabier ,
Marc Pollefeys ,
Martin R. Oswald
ICCVW , 2019
arXiv
Our method learns sensor or algorithm properties jointly with semantic depth fusion and scene completion and can also be used as an expert system, eg to unify the strengths of various photometric stereo algorithms.
Your browser does not support the video tag.
The World of Fast Moving Objects
Denys Rozumnyi ,
Jan Kotera ,
Filip Sroubek ,
Lukas Novotny ,
Jiri Matas
CVPR , 2017
project page /
arXiv /
poster /
poster source
Introducing fast moving objects for the first time as objects that move over distances larger than their size in one video frame: new problem, new dataset, new metrics, new baseline.
Coplanar Repeats by Energy Minimization
James Pritts ,
Denys Rozumnyi ,
M. Pawan Kumar ,
Ondřej Chum
BMVC , 2016
arXiv
We propose an automated method to detect, group and rectify arbitrarily-arranged coplanar repeated elements via energy minimization.
Supervising
Francesco Girlanda Zou : Gaussian Splatting SLAM with Deblurring Capabilities , Master thesis, ongoing.
Rong Zou : Retrieval Robust to Object Motion Blur , Master thesis, Nov 2023, accepted to ECCV'24 .
Yiming Zhao : Human from Blur: Human Pose Tracking from Blurry Images , Semester project, May 2023, accepted to ICCV'23 .
Rajat Thakur : Predicting 3D Shape and Texture of Fast Moving Cars , Semester project defended in March 2022.
Adrian Klaeger : Temporal Super-Resolution of Multiple Fast-Moving Objects , Master thesis defended in September 2021, thesis , GitHub .
Harish Rajagopal : Improving DeFMO With Learned Losses , Semester project defended in June 2021, report , GitHub .
Julius Fricke : ADMM Algorithm Unrolling: Deblurring and Matting , Bachelor thesis defended in April 2021.
Reviewing
Conferences : CVPR, ICCV, ECCV, NeurIPS, ICML, ICLR, 3DV, WACV.
Journals : PAMI, IJCV.