Search
Search
AI Chat
Recommendation
Reviewer
PDF Analysis
Projects
Sign in

Sign in to unlock AI Chat and more features

DINOv2: Learning Robust Visual Features without Supervision

arXiv (Cornell University)

Maxime Oquab, Timothée Darcet, Théo Moutakanni, Huy Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mahmoud Assran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Michael Rabbat, Vasu Sharma, Gabriel Synnaeve, Hu Xu, Hervé Jegou, Julien Mairal, Patrick Labatut, Armand Joulin, Piotr Bojanowski

Citations:

627

References:

131

Abstract
The recent breakthroughs in natural language processing for model pretraining on large quantities of data have opened the way for similar foundation models in computer vision. These models could greatly simplify the use of images in any system by producing all-purpose visual features, i.e., features that work across image distributions and tasks without finetuning. This work shows that existing pr...
See More
View PDF
Keywords
Computer science
Pipeline (software)
Artificial intelligence
Machine learning
Scale (ratio)
Training set
Image (mathematics)
Physics
Quantum mechanics
Programming language
Sustainable Development Goals (SDG)
Quality education
2023
Preprint

Open Access

Institution

Inria. Théo Moutakanni has, 

co-affiliation with Université Paris Saclay. Alaaeldin

Publisher

Cornell University

2-Y Citation Average

0.493

DOI

https://doi.org/10.48550/arxiv.2304.07193

Language

en

Country

France

pdf file
DINOv2: Learning Robust Visual Features without Supervision
pdf file

DINOv2: Learning Robust Visual Features without Supervision

arXiv (Cornell University)

Maxime Oquab, Timothée Darcet, Théo Moutakanni, Huy Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mahmoud Assran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Michael Rabbat, Vasu Sharma, Gabriel Synnaeve, Hu Xu, Hervé Jegou, Julien Mairal, Patrick Labatut, Armand Joulin, Piotr Bojanowski

Citations:

627

References:

131

Abstract
The recent breakthroughs in natural language processing for model pretraining on large quantities of data have opened the way for similar foundation models in computer vision. These models could greatly simplify the use of images in any system by producing all-purpose visual features, i.e., features that work across image distributions and tasks without finetuning. This work shows that existing pr...
See More
View PDF
Keywords
Computer science
Pipeline (software)
Artificial intelligence
Machine learning
Scale (ratio)
Training set
Image (mathematics)
Physics
Quantum mechanics
Programming language
Sustainable Development Goals (SDG)
Quality education
2023
Preprint

Open Access

Institution

Inria. Théo Moutakanni has, 

co-affiliation with Université Paris Saclay. Alaaeldin

Publisher

Cornell University

2-Y Citation Average

0.493

DOI

https://doi.org/10.48550/arxiv.2304.07193

Language

en

Country

France

pdf file