Iaroslav (Iaro) Melekhov

CV   MAIL   GSCHOLAR   GITHUB   LINKEDIN   TWITTER

Computer vision researcher at Aalto University.

During my PhD, I was supervised by Prof. Juho Kannala and Prof. Esa Rahtu. My research interests lie at the intersection of 3D Computer Vision, Scene Understanding, and Deep Learning.

Previously, I was a research intern at Niantic Labs (London) under the supervision of Prof. Gabriel Brostow and Dr. Daniyar Turmukhambetov. Before this, I spent 6 months in Cambridge, UK working on visual perception algorithms for autonomous cars and having fun at Wayve. I also did a research visit to ETH Zurich's CVG Group led by Prof. Marc Pollefeys.

News

11/2023: Our work on visual localization has been accepted to IJCV!
10/2021: One paper got accepted to 3DV 2021!
07/2021: Our paper Continual Learning for Image-Based Camera Localization got accepted to ICCV 2021.
06/2021: I started as a Postdoctoral researcher under the supervision of Jaakko Lehtinen here at Aalto.
08/2020: Our new approach secured 2nd place in two ECCV 2020 competitions on image-based localization (1 and 2). The slides are available online and this is a project I did during my internship at Niantic Labs.
03/2020: I successfully defend my PhD thesis on image matching and camera relocalization 🎉🎉🎓! I am very thankful to the thesis pre-examiners (Dr. Vassileios Balntas from FAIR and Prof. Atsuto Maki from KTH) for their comments and very useful feedback and to my opponent (Prof. Victor Lempitsky from Samsung AI Center) for the great discussion during defense.
06/2019: I am very excited to announce that I started a 4-month research internship at Niantic Labs, London.

Research

Digging Into Self-Supervised Learning of Feature Descriptors

Iaroslav Melekhov*, Zakaria Laskar*, Xiaotian Li, Shuzhe Wang, Juho Kannala

3DV 2021

paper   webpage   code   bibtex
@inproceedings{Melekhov2021hndesc,
title = {Digging Into Self-Supervised Learning of Feature Descriptors},
author = {Melekhov, Iaroslav and Laskar, Zakaria and Li, Xiaotian and Wang, Shuzhe and Kannala Juho},
booktitle = {In Proceedings of the International Conference on 3D Vision (3DV)},
year = {2021}}

@article{Melekhov2020Nian,
author = {{Melekhov, Iaroslav and Brostow, Gabriel J. and Kannala, Juho and Turmukhambetov, Daniyar},
title = {Image Stylization for Robust Features},
journal = {Arxiv preprint arXiv:2008.06959},
year = {2020}}

Continual Learning for Image-based Camera Localization

Shuzhe Wang*, Zakaria Laskar*, Iaroslav Melekhov, Xiaotian Li, Juho Kannala

ICCV 2021

paper   code   bibtex
@inproceedings{Wang2021clloc,
title = {Continual Learning for Image-based Camera Localization},
author = {Wang, Shuzhe and Laskar, Zakaria and Melekhov, Iaroslav and Li, Xiaotian and Kannala Juho},
booktitle = {In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
year = {2021},
}

Image Stylization for Robust Features

Iaroslav Melekhov, Gabriel Brostow, Juho Kannala, Daniyar Turmukhambetov

Paper preprint, 2020

paper   code   bibtex
@article{Melekhov2020Nian,
author = {{Melekhov, Iaroslav and Brostow, Gabriel J. and Kannala, Juho and Turmukhambetov, Daniyar},
title = {Image Stylization for Robust Features},
journal = {Arxiv preprint arXiv:2008.06959},
year = {2020}}

Geometric Image Correspondence Verification by Dense Pixel Matching

Zakaria Laskar*, Iaroslav Melekhov*, Hamed R. Tavakoli, Juha Ylioinas, Juho Kannala

WACV 2020

paper   bibtex
@InProceedings{LaskarWACV2020,
author = {Laskar, Zakaria and Melekhov, Iaroslav and Tavakoli, Hamed Rezazadegan and Ylioinas, Juha and Kannala, Juho},
title = {Geometric Image Correspondence Verification by Dense Pixel Matching},
booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
month = {March},
year = {2020},
}

TC-Net: Self-Supervised Monocular Video Scene Understanding Using Temporally Consistent Geometric Prior

Iaroslav Melekhov, Esa Rahtu, Juho Kannala, Alex Kendall

ICML Workshops 2019

paper bibtex
@online{MelekhovICMLW19,
author = {Melekhov, Iaroslav and Rahtu, Esa and Kannala, Juho and Kendall, Alex},
title = {{TC-Net: Self-Supervised Monocular Video Scene Understanding Using Temporally Consistent Geometric Prior}},
date = {2019},
annotation = {International Conference on Machine Learning (ICML) Workshops},
}

DGC-Net: Dense Geometric Correspondence Network

Iaroslav Melekhov, Aleksei Tiulpin, Torsten Sattler, Marc Pollefeys, Esa Rahtu, Juho Kannala

WACV 2019, best paper honorable mention award

paper   code   project page   video   bibtex
@inproceedings{MelekhovWACV2018,
title = {{DGC-Net}: Dense geometric correspondence network},
author = {Melekhov, Iaroslav and Tiulpin, Aleksei and Sattler, Torsten, and Pollefeys, Marc and Rahtu, Esa and Kannala, Juho},
year = {2019},
booktitle = {Proceedings of the IEEE Winter Conference on Applications of Computer Vision (WACV)},
}

Image-based Localization using Hourglass Networks

Iaroslav Melekhov, Juha Ylioinas, Juho Kannala, Esa Rahtu

ICCV Workshops 2017

paper   bibtex
@InProceedings{MelekhovICCVW2017,
author = {Melekhov, Iaroslav and Ylioinas, Juha and Kannala, Juho and Rahtu, Esa},
title = {Image-Based Localization Using Hourglass Networks},
booktitle = {Proceedings of the IEEE International Conference on Computer Vision (ICCV) Workshops},
month = {Oct},
year = {2017},
}

Camera Relocalization by Computing Pairwise Relative Poses Using Convolutional Neural Network

Zakaria Laskar*, Iaroslav Melekhov*, Surya Kalia, Juho Kannala

ICCV Workshops 2017

paper   code   bibtex
@InProceedings{Laskar_2017_ICCV_Workshops,
author = {Laskar, Zakaria and Melekhov, Iaroslav and Kalia, Surya and Kannala, Juho},
title = {Camera Relocalization by Computing Pairwise Relative Poses Using Convolutional Neural Network},
booktitle = {Proceedings of the IEEE International Conference on Computer Vision (ICCV) Workshops},
month = {Oct},
year = {2017},
}

Relative Camera Pose Estimation Using Convolutional Neural Networks

Iaroslav Melekhov, Juha Ylioinas, Juho Kannala, Esa Rahtu

ACIVS 2017

paper   code   bibtex  
@InProceedings{MelekhovACIVS2017,
author = {Melekhov, Iaroslav and Ylioinas, Juha and Kannala, Juho and Rahtu, Esa},
title = {Relative Camera Pose Estimation Using Convolutional Neural Networks},
booktitle = {Advanced Concepts for Intelligent Vision Systems (ACIVS)},
year = {2017} }

Siamese Network Features for Image Matching

Iaroslav Melekhov, Juho Kannala, Esa Rahtu

ICPR 2016

paper   bibtex  
@InProceedings{MelekhovICPR2016,
author={I. {Melekhov} and J. {Kannala} and E. {Rahtu}},
title = {Relative Camera Pose Estimation Using Convolutional Neural Networks},
booktitle={2016 23rd International Conference on Pattern Recognition (ICPR)},
year = {2016},
}

Image Patch Matching Using Convolutional Descriptors with Euclidean Distance

Iaroslav Melekhov, Juho Kannala, Esa Rahtu

ACCV Workshops 2016

paper   bibtex  
@InProceedings{MelekhovACCVW2016,
author={I. {Melekhov} and J. {Kannala} and E. {Rahtu}},
title = {Image Patch Matching Using Convolutional Descriptors with Euclidean Distance},
booktitle={Asian Conference on Computer Vision (ACCV) Workshops},
year = {2016},
}

Talks

Aug. 2020: Image stylization for robust features. ECCV 2020 Workshop on Long-Term Visual Localization under Changing Conditions. Glasgow, UK.
link   slides (24MB)
Aug. 2020: Image stylization for robust features. ECCV 2020 Workshop on Map-based Localization for Autonomous Driving. Glasgow, UK.
link   slides (24MB)