ECCV 2018 Tutorial

Visual Localization

When: Saturday, September 8th, 2018
Where: N1070ZG, Technische Universität München Arcisstraße 21, 80333 Munich (https://goo.gl/maps/tZENRFMpcwD2)
Time: Morning
Schedule
- 9:00 - 9:10 Introduction [10 min] (Eric, Torsten) [slides][slides with intermediate stages]
- 9:10 - 10:00 Current State of Feature-based Localization [50 min] (Torsten) [slides][slides with intermediate stages]
  - Basics: Local features, camera pose estimation, descriptor matching [10 min]
  - Efficient (Mobile) Localization [20 min]
  - Scalable Feature-based Localization [20 min]
- 10:00 - 10:15 Coffee Break
- 10:15 - 11:15 Current State of Learning-based Localization [60 min] (Eric) [slides]
  - Basics: Random Forests, Convolutional Neural Networks [10 min]
  - Learning Camera Pose Regression [20 min]
  - Scene Coordinate Regression [30 min]
- 11:15 - 11:30 Coffee Break
- 11:30 - 12:20 Current Topics & Open Problems [50 min] (Eric, Torsten)
  - Failure Cases of Feature-based and Learning-based Methods [10 min] [slides][slides with intermediate stages]
  - Long-term Localization: Towards Higher-level Scene Understanding [20 min] [slides][slides with intermediate stages]
  - Open Problems of Learning-based Methods [20 min] [slides]
- 12:20 - 12:40 Questions & Discussion [20 min]

Please see https://sites.google.com/view/lsvpr2017/home for pointers to publicly available code and publicly available datasets.