This praktikum introduces common tools and principles in computer vision and machine learning for object recognition and pose estimation, for applications in the field of Augmented Reality and Robotics. Starting from the basics, students are gradually taught to develop a framework that can process images and 3D data with the goal of perceiving shapes and recognizing objects in real environments, under pose variations, clutter and occlusion.
Lecturers: Dr. Federico Tombari, Prof. Nassir Navab| Date | Location | Topic | Assignments | Conducted by | Materials | Due |
|---|---|---|---|---|---|---|
| 02.02.2017, 18:00 | 03.13.010 | Preliminary meeting ("Vorbesprechung") | - | Course Tutors | Course info | - |
| 27.04.2017, 14.00 - 15.30 | 02.07.023 | Introduction & Computer Vision basics | OpenCV | Federico | Lecture Exercise Code (fixed) | 03.05.2017 23:59 |
| 04.05.2017, 14.00 - 15.30 | 02.07.023 | Feature description and matching | OpenCV | Federico | Lecture Exercise Code | 10.05.2017 23:59 |
| 11.05.2017, 14.00 - 15.30 | 02.07.023 | Fundamentals of CNNs and deep feature learning | TensorFlow | Iro | Lecture Exercise Code | 18.05.2017 23:59 |
| 18.05.2017, 14.00 - 15.30 | 02.07.023 | Image classification and object detection with CNNs | TensorFlow | Iro | Lecture Exercise Code | 30.05.2017 23:59 |
| 01.06.2017, 14.00 - 15.30 | 02.07.023 | 3D sensors and 3D representations | OpenNI, Eigen | Federico | Lecture Exercise Code | 07.06.2017 23:59 |
| 08.06.2017, 14.00 - 15.30 | 02.07.023 | Surface matching and registration | PCL | Federico | Lecture Exercise Code | 21.06.2017 23:59 |
| 22.06.2017, 14.00 - 15.30 | 02.07.023 | 3D descriptors for object detection and 6DoF pose estimation | PCL | Federico | Lecture Exercise Code | 28.06.2017 23:59 |
| 29.06.2017, 14.00 - 15.30 | 02.07.023 | Material for Challenge I | Federico, Christian | Slides Validation Data | - | |
| 06.07.2017, 14.00 - 15.30 | 02.07.023 | Support for Challenge I | Federico, Christian | - | 13.07.2017 23:59 | |
| 13.07.2017, 14.00 - 15.30 | 02.07.023 | Material for Challenge II | Federico, Christian | Yet Another Training Dataset Challenge2 Validation | cancelled,used for final challenge only | |
| 03.08.2017, 14.00 - 18.00 | 00.13.009A | Final Challenge | Final Challenge Images Secret Object |
| TeachingForm | |
|---|---|
| Title: | Perception and Learning in Robotics and Augmented Reality |
| Professor: | Dr. Federico Tombari, Prof. Nassir Navab |
| Tutors: | Christian Rupprecht, Iro Laina; Keisuke Tateno |
| Type: | Praktikum |
| Information: | 6 SWS, 10 ECTS Credits (IN2106) |
| Term: | 2017SoSe |
| Abstract: | This praktikum introduces common tools and principles in computer vision and machine learning for object recognition and pose estimation, for applications in the field of Augmented Reality and Robotics. Starting from the basics, students are gradually taught to develop a framework that can process images and 3D data with the goal of perceiving shapes and recognizing objects in real environments, under pose variations, clutter and occlusion. |
