Robust and Efficient Perception for Autonomous Things (REPEAT)

Posted on 24/09/19.

One of the grand challenges in artificial intelligence is to build machines that are able to perceive and navigate in complex environments. Typical applications of such systems include autonomous vehicles, service and manufacturing robots, drones, and augmented reality devices. The purpose of this project is to study computer vision methods for odometry, simultaneous localisation and mapping (SLAM), image based localization, 3D reconstruction, and object detection and pose estimation, which are all key components in autonomous operation of mobile machines.

The emphasis will be on the development of deep learning methodology in the aforementioned problems. One specific object of study is deep learning based multi-view 3D reconstruction from videos. This work has potential to impact robotics and autonomous driving by decreasing dependency on expensive and heavy lidar sensors.

Robust and Efficient Perception for Autonomous Things (REPEAT)

Project leader: Esa Rahtu, Tampere University
Project coordinator at the University of Vaasa: Jani Boutellier
Time: 1.1.2020-31.12.2022
Budget: 326 028 euros (University of Vaasa) / 983 752 euroa (consortium)
Funding from: Academy of Finland
External funding: 228 220 euros (97 808 euros for the University of Vaasa)
Contact person at the University of Vaasa: Jani Boutellier
Research platform: Digital Economy
Project partners: University of Vaasa, Aalto University, Tampere University

Project web site: https://sites.univaasa.fi/repeat/

Back