Workshop on Multimodal Analysis of Spatio-Temporal Data
The workshop is organized within the BMBF project MM4SPA - Multimodal Analysis for Sports Analytics (duration: 2020-2023) which is a collaboration of the German Sport University Cologne, Leibniz University Hannover and Kinexon GmbH. In the project, state-of-the-art approaches for intelligent synchronization, semantic enrichment, and analysis of sports data from soccer and handball are being developed. The multimodal information from video recordings, position data (video- or sensor-based tracking) as well as manual annotations by experts are used for the recognition and evaluation of tactical movement patterns, sport-specific events and game performance.
The aim of this workshop is interdisciplinary networking and exchange between researchers who analyze spatio-temporal data or multimodal information gained from different sensor information like videos and GPS coordinates. With advances in sensor technology and computer vision, the analysis of movement patterns using semantically enriched motion data from different sensor sources has found its way into various research fields such as computer science, sports science, geosciences, biology, transportation, and logistics.
Based on presentations from researchers and experts from different domains, this workshop aims to discuss open research questions, proposed solutions as well as their limitations regarding the multimodal use of sensor data:
- data sources, quality, characteristics, and preprocessing
- state-of-the-art approaches for the analysis of spatio-temporal data
- domain-specific issues for the analysis of spatio-temporal data
Venue
The workshop will be held in a hybrid format, i.e., it will take place both online and on-site at Appelstraße 9a, 15th floor of the L3S Research Center at Leibniz University Hannover, Hannover, Germany.
Date: 28th March 2023
Time: 10 00 am - 4 30 pm (CET)
Venue: Appelstraße 9a, Hannover & online
Registration
For registration please write an email to and indicate if you will participate on site or virtually.
An invitation link for virtual participation will be shared later by mail.
Schedule
The workshop is divided into two sessions:
- Session 1: four keynote presentations from different research areas with Q&A
- Session 2: results and insights from the MM4SPA project
- four research talks on the analysis of sport games using position and video data
- one keynote on the findings and implications for the industry partner Kinexon
| Time (CET) | Speaker | Title | Presentation |
|---|---|---|---|
| 10 00 - 10 15 | Daniel Memmert & Ralph Ewerth | Opening | |
| 10 15 - 10 55 | Alexander Dockhorn | Generierung Prädiktiver Modelle in unbekannten Spielen | 30 min Keynote + 10 min Q&A |
| 11 00 - 11 40 | Prasenjit Mitra | TBA | 30 min Keynote + 10 min Q&A |
| 11 45 - 12 25 | Andrea Flack | Using Tracking Devices to Understand Stork Migration | 30 min Keynote + 10 min Q&A |
| 12 30 - 13 10 | TBA | TBA | 30 min Keynote + 10 min Q&A |
| Lunch | |||
| 14 00 - 14 20 | Henrik Biermann & Manuel Bassek, Robert Rein, Daniel Memmert (DSHS Köln) | Contextualization of Analysis in Handball and Soccer with Tactical Periodization | 15 min + 5 min Q&A |
| 14 25 - 14 45 | Jonas Theiner (L3S) | Sports Field Registration as Multimodal Fusion of Player Position and Video Data | 15 min + 5 min Q&A |
| 14 50 - 15 10 | Dominik Raabe, Robert Rein, Daniel Memmert (DSHS Köln) | Spatiotemporal Data Analysis with Graph-based Deep Learning | 15 min + 5 min Q&A |
| 15 15 - 15 35 | TBA (L3S) | TBA | 15 min + 5 min Q&A |
| 15 40 - 16 25 | Philipp Seidenschwarz (Kinexon) | Practical Implications from Kinexon | 30 min Keynote + 10 min Q&A |
| 16 20 | Closing | ||
| 16 30 - ? | Come-together |
Organizers
- Prof. Dr. Ralph Ewerth (L3S Research Center, Leibniz Universität Hannover & TIB - Leibniz Information Centre for Science and Technology, Hannover, Germany)
- Prof. Dr. Daniel Memmert (Institute of Exercise Training and Sport Informatics, German Sport University Cologne)