Contact

Karlovo namesti 13, 121 35 Prague, Czech Republic (map)
Room: E227
E-mail: vacekpa2@fel.cvut.cz

Publications

• Learning to predict lidar intensities. P. Vacek, O. Jasek, K. Zimmermann, T. Svoboda - IEEE Transactions on Intelligent Transportation Systems, 10.1109/TITS.2020.3037980. April 2022.

• Real3D-Aug: Point Cloud Augmentation by Placing Real Objects with Occlusion Handling for 3D Detection and Segmentation. Petr Šebek, Šimon Pokorný, Patrik Vacek, Tomáš Svoboda. 2022. Computer Vision Winter Workshop 2023 Arxiv

• Teachers in concordance for pseudo-labeling of 3D sequential data. Awet Haileslassie Gebrehiwot, Patrik Vacek, David Hurych, Karel Zimmermann, Patrick Perez, Tomáš Svoboda - IEEE Robotics and Automation Letters. 2022 Arxiv

Collaboration

Valeo R&D center in Prague and Valeo.ai

Interests and research

I have studied deep learning methods in various computer vision tasks, especially in robotics and autonomous driving. I focus on semi-supervised and weakly-supervised learning for 3D point clouds. Currently, my interest lies in supervision from ego odometry for unsupervised learning of detections during driving and supervision from physical constraints such as motion consistency.

Teaching

• B3B33VIR - Vision For Robotics (Vidění robotu) (Winter Semester 2019-2022)

Github