Hauptseminar/Master Seminar:
Recent Trends in 3D Computer Vision and Deep Learning

Prof. Nassir Navab, Federico Tombari, Iro Laina, Christian Rupprecht, Helisa Dhamo

Course number: Module IN4826
Type: Advanced Seminar Course Module IN8901
Type: Master Seminar Module IN2107
Type: Seminar Module IN0014
SWS: 2+0
ECTS: 5 Credits
Location: TBD
Time: TBD
Course Language: English

Announcements

• The preliminary meeting was on July 6, 16:00, Location: 02.08.011
  
  

Introduction

• The Seminar Course concerns recent advances in the field of 3D Computer Vision and Deep Learning. The Seminar will propose a list of recent scientific articles related to the main current research topics in the field, such as 3D keypoint detection and description, point cloud matching and registration, 3D object recognition and classification, SLAM, 3D reconstruction, 3D object pose estimation and tracking, depth prediction, 3D scene understanding.

Registration

• Preliminary Meeting: Slides
• The registration will be performed through the TUM Matching Platform.

Requirements

In this Master Seminar (formerly Hauptseminar), each student is asked to select one paper from a list. In order to successfully complete the seminar, participants have to fulfill these requirements:

• Presentation: The selected paper is presented to the other participants (20 minutes presentation 10 minutes questions). Use the CAMP templates for PowerPoint camp-tum-jhu-slides.zip, or Latex: CAMP-latex-template.
• Written Report: A document of maximum 8 pages is written and submitted one week after the presentation. Please download and use the llncs2e.zip Latex template.
• Attendance: Participants have to participate actively in all seminar sessions.

The students are required to attend each seminar presentation which will be held during this course. Each presentation is followed by a discussion and everyone is encouraged to actively participate. The report must include all references used and must be written completely in your own words. Copy and paste will not be tolerated. Both report and presentation have to be done in English.

List of Topics and Material

Preliminary list of papers. Subject to changes.

Date	Authors	Title	Conference	Link	Tutor	Student
Mi 18.10.2017	J. Martinez	On human motion prediction using recurrent neural networks	CVPR 2017	link	Huseyin	Kankamon Rujiranun
Mi 18.10.2017 Mi 29.11.2017	A. Alahi et al.	Social LSTM: Human Trajectory Prediction in Crowded Spaces	CVPR 2016	link	Oliver	Thomas Baumeister
Mi 08.11.2017	Wang et al.	SURGE: Surface Regularized Geometry Estimation from a Single Image	NIPS 2016	link	Iro	Michael Lohr
Mi 08.11.2017	Zhou et al.	Unsupervised Learning of Depth and Ego-Motion from Video	CVPR 2017	link	Iro	Thomas Schmid
Mi 22.11.2017	R. Mur-Artal and J. D. Tardos	Visual-Inertial Monocular SLAM with Map Reuse	RAL 2017	link	Chen Xiao	Robin Ostner
Mi 22.11.2017	K. Tateno et al.	CNN-SLAM: Real-time dense monocular SLAM with learned depth prediction	CVPR 2017	link	Yida	Florian Windolf
Mi 29.11.2017	T. Cavallari et al.	On-the-Fly Adaptation of Regression Forests for Online Camera Relocalisation	CVPR 2017	link	Helisa	Christoph Gnoth
Mi 29.11.2017	J. Engel et al.	Direct sparse odometry	PAMI 2017	link	Jakob	Vera Hug
Mi 06.12.2017	G. Riegler et al	OctNetFusion: Learning Depth Fusion from Data	3DV 2017	link	Keisuke	Rupal Jain
Mi 06.12.2017 Mi 10.01.2018	M. Slavcheva et at.	KillingFusion: Non-rigid 3D Reconstruction without Correspondences	CVPR 2017	link	Patrick	Mohammad Ahantab
Mi 13.12.2017	C.R. Qi et al.	PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space	NIPS 2017	link	Johanna	Johann Frei
Mi 13.12.2017	J. Wu et al.	Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling	NIPS 2016	link	Niko	Jonas Gerstner
Mi 10.01.2017	N. Verma et al.	Dynamic Filters in Graph Convolutional Networks	arxiv 2017	link	Tolga	Anna Reithmeir
Mi 10.01.2017	F. Monti et al.	Geometric Deep Learning on Graphs and Manifolds Using Mixture Model CNNs	CVPR 2017	link	Benni	Tobias Valinski
Mi 24.01.2018	M. Rahd and V. Lepetit	BB8: A Scalable, Accurate, Robust to Partial Occlusion Method for Predicting the 3D Poses of Challenging Objects without Using Depth	arxiv 2017	link	Fabian	Saadhana Boppineni Venkataraman
Mi 24.01.2018	Haoqiang Fan et. al.	A Point Set Generation Network for 3D Object Reconstruction from a Single Image	arXiv 2017	link	Christian	Alexei Vasilev

Schedule

The schedule for the paper presentations will be announced here during the semester. Each student is required to attend all presentations.

Date	Time	Place	Speaker	Topic
Mi 18.10.2017	16:00 - 17:30	03.13.010	Kankamon Rujiranun	Human motion prediction
Mi 08.11.2017	16:00 - 17:30	03.13.010	Michael Lohr Thomas Schmid	Depth estimation
Mi 15.11.2017	16:00 - 17:30	03.13.010		cancelled
Mi 22.11.2017	16:00 - 17:30	03.13.010	Robin Ostner Florian Windolf	Simultaneous localization and mapping
Mi 29.11.2017	16:00 - 17:30	03.13.010	Christoph Gnoth Vera Hug Thomas Baumeister	Localization
Mi 06.12.2017	16:00 - 17:30	03.13.010	Rupal Jain	Reconstruction
Mi 13.12.2017	16:00 - 17:30	03.13.010	Johann Frei Jonas Gerstner	Learning geometry
Mi 10.01.2018	16:00 - 17:30	03.13.010	Anna Reithmeir Tobias Valinski Mohammad Ahantab	Graph convolutional networks
Mi 24.01.2018	16:00 - 17:30	03.13.010	Saadhana Boppineni Venkataraman Alexei Vasilev	Learning with objects