Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
One net to rule them all (global localization and pose estimation)
Combine state-of-the-art place embedding and feature extraction networks into a single network and train them jointly.
Keywords: Localization place recognition CNN computer vision
Global localization and (local) pose estimation are complementary tasks, see https://youtu.be/c0LwfDJ_F2A?t=52 . We would like to investigate baking them both into a single network, jointly trained.
Global localization and (local) pose estimation are complementary tasks, see https://youtu.be/c0LwfDJ_F2A?t=52 . We would like to investigate baking them both into a single network, jointly trained.
Combine state-of-the-art place embedding and feature extraction networks into a single network and train them jointly. We believe this can be done in a simpler way than what is currently proposed in related work.
Combine state-of-the-art place embedding and feature extraction networks into a single network and train them jointly. We believe this can be done in a simpler way than what is currently proposed in related work.
Titus Cieslewski ( titus at ifi.uzh.ch ), ATTACH CV AND TRANSCRIPT (also Bachelor)! Preferred skills: Linux, Python, some Computer Vision background, TensorFlow/PyTorch or equivalent. This project will be co-supervised by Dimche Kostadinov.
Titus Cieslewski ( titus at ifi.uzh.ch ), ATTACH CV AND TRANSCRIPT (also Bachelor)! Preferred skills: Linux, Python, some Computer Vision background, TensorFlow/PyTorch or equivalent. This project will be co-supervised by Dimche Kostadinov.