I am a Doctoral candidate in computer vision at Aalto University under the supervision of Dr. Juho Kannala and Dr. Esa Rahtu.
During my PhD, I have been focusing on closely related computer vision problems, such as camera relocalisation (image-based localisation), image similarity, dence correspondence search, and applying Machine Learning to these tasks. At the beginning of October, I will start a 5-month research internship (until March 2019) at wayve.ai in Cambridge, UK with the focus on scene-understanding for autonomous vehicles.
Visit my google scholar profile.
- I. Melekhov, E. Rahtu, J. Kannala, A. Kendall, TC-Net: Self-Supervised Monocular Video Scene Understanding Using Temporally Consistent Geometric Prior, ICML-SSL 2019
- I. Melekhov, A. Tiulpin, T. Sattler, M. Pollefeys, E. Rahtu, J. Kannala, DGC-Net: Dense Geometric Correspondense Network, WACV 2019 Best paper honorable mention award
[pdf] [project page]
- Z. Laskar*, I. Melekhov*, S. Kalia, J. Kannala, Camera Relocalization by Computing Pairwise Relative Pose using Convolutional Neural Network, ICCV-GMDL 2017 (oral)
[pdf] [code] [poster]
- I. Melekhov, J. Ylioinas, J. Kannala, E. Rahtu, Image-based Localization using Hourglass Networks, ICCV-GMDL 2017
- I. Melekhov, J. Ylionas, J. Kannala, E. Rahtu, Relative Camera Pose Estimation using Convolutional Neural Networks, ACIVS 2017 (oral)
[pdf] [code] [slides]
- I. Melekhov, J. Kannala, E. Rahtu, Siamese Network Features for Image Matching, ICPR 2016
- I. Melekhov, J. Kannala, E. Rahtu, Image Patch Matching using Convolutional Descriptors with Euclidean Distance, ACCV-IVDNN 2016 (oral)
iaroslav dot melekhov at aalto dot fi
github Linkedin profile twitter