Research

Human activity is considered to be one of the most important ingredients of context awareness and ubiquitous computing. During my PhD I focused mainly on the recognition complex activities based on wearable sensors. By spotting isolated gesture-alike events in a continuous data stream, probabilistic temporal models are used to compose such events into higher level activities such as daily routines, or multi-step construction tasks.
"We are what we repeatly do." (Aristotle, 384-322 BC)
Dissertation
Recognizing Complex Human Activity Based on Activity Spotting Ulf Blanke Prof. Stefan Roth, Ph. D (examiner) Prof. Dr.-Ing Hans Gellersen (co-examiner) Prof. Dr. Bernt Schiele (co-examiner) Dissertation (defended 12/2010)
Peer reviewed publications
2014
Crowdsourced Pedestrian Map Construction for Short-Term City-Scale Events U. Blanke, R. Robin, S. Feese, G. Tröster The First International Conference on IoT in Urban Space (Urb-IOT) In-cooperation with ACM SIGAPP, SIGCHI, and SIGSPATIAL, and Technically Co-Sponsored by IEEE ComSoC IoT TC Rome, Italy
Inferring Travel Purpose from Crowd-Augmented Human Z. Zhu, U. Blanke, G. Tröster The First International Conference on IoT in Urban Space (Urb-IOT) In-cooperation with ACM SIGAPP, SIGCHI, and SIGSPATIAL, and Technically Co-Sponsored by IEEE ComSoC IoT TC Rome, Italy
Participatory Bluetooth Scans Serving as Urban Crowd Probes J. Weppner, P. Lukowicz, U. Blanke, G. Tröster IEEE Sensors Journal Special Issue on Sensing technologies for intelligent urban infrastructures 2014
How to Save Human Lives with Complexity Science D.Helbing, D. Brockmann, T. Chadefaux, K. Donnay, U. Blanke, O. Woolley-Meza, M.Moussaid, A. Johansson, J. Kraus, S.Schutte and M. Perc Social Science Research Network (preprint)
Capturing crowd dynamics at large scale events using participatory GPS-localization U. Blanke, T. Franke, G. Tröster, P. Lukowicz The 2014 IEEE Ninth International Conference on Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP) Singapore.
GaitAssist: A Daily-Life Support and Training System for Parkinson’s Disease Patients with Freezing of Gait S. Mazilu, U. Blanke, M. Hardegger, E. Gazit, J. M. Hausdorff, G. Tröster ACM CHI Conference on Human Factors in Computing Systems. (CHI) Toronto Oral, 22.8% acceptance rate
(Demo) GaitAssist: A Wearable Assistant for Gait Training and Rehabilitation in Parkinson's Disease S. Mazilu, M. Hardegger, U.Blanke, G. Tröster 12th IEEE Intl. Conference on Pervasive Computing and Communications (PerCom). Budapest, Hungary. Best demo award
The Telepathic Phone: Frictionless Activity Recognition from WiFi-RSSI S. Sigg, U.Blanke, G. Tröster 12th IEEE Intl. Conference on Pervasive Computing and Communications (PerCom). Budapest, Hungary. Oral, 14% acceptance rate
A Tutorial on Human Activity Recognition Using Body-worn Inertial Sensors A. Bulling, U. Blanke, and B. Schiele. ACM Computing Surveys * in equal contribution
2013
Mobile Computing, Applications, and Services G. Memmi (General Chair) U. Blanke (TPC Chair) 5th International Conference, MobiCase 2013, Paris, France, November 7-8, 2013, Revised Selected Papers
Human Activity Recognition Using Social Media Data Z. Zhu, U.Blanke, A. Calatroni, G. Tröster 12th Intl. ACM Conference on Mobile and Ubiquitous Multimedia (MuM) Lulea, Sweden. Best paper award
Towards Scalable Activity Recognition: Adapting Zero-Effort Crowdsourced Acoustic Models LV. Nguyen-Dinh, U. Blanke, G. Tröster 12th Intl. ACM Conference on Mobile and Ubiquitous Multimedia (MuM) Lulea, Sweden.
Low-Power Ambient Sensing in Smartphones for Continuous Semantic Localization S. Mazilu, U. Blanke, A. Calatroni, G. Tröster 4th International Joint Conference on Ambient Intelligence (AmI) Dublin, Ireland, 2013
Combining Crowd-Generated Media and Personal Data: Semi-Supervised Learning for Context Recognition L.V. Nguyen-Dinh, M. Rossi, U. Blanke and G. Tröster ACM Workshop on Personal Data Meets Distributed Multimedia co-located with ACM Multimedia, Barcelona, Spain, 2013.
Prior Knowledge of Human Activities from Crowd-Generated Social Data Z. Zhu, U. Blanke, G. Tröster. 17th ACM/IEEE Intl. Symposium on Wearable Computing 2013. (ISWC). Zürich, Switzerland.
Engineers Meet Clinicians: Augmenting Parkinson's Disease Patients to Gather Information for Gait Rehabilitation S. Mazilu, E. Gazit, U. Blanke, D. Roggen, J. Hausdorff and G. Tröster. ACM SIGCHI Augmented Human 2013. (AH) Stuttgart, Germany..
User Independent, Multi-Modal Spotting of Subtle Arm Actions with Minimal Training Data. G. Bauer, P. Lukowicz, U. Blanke, B. Schiele. 10th IEEE Percom Workshop on Context Modeling and Reasoning 2013 (CoMoRea). San Diego
2012
Ah! I felt like doing something else...
2011
I spy with my little eye: Learning Optimal Filters for Cross-Modal Stereo under Projected Patterns Walon Chiu, Ulf Blanke, Mario Fritz Max Planck Institute for Informatics 1st IEEE Workshop on Consumer Depth Cameras for Computer Vision in conjuction with ICCV Barcelona, Spain.
Improving the Kinect by Cross-Modal Stereo Walon Chiu, Ulf Blanke, Mario Fritz Max Planck Institute for Informatics The 22nd British Machine Vision Conference (BMVC), Dundee, UK.
South by South-East Or Sitting at the Desk. Can Orientation be a place? Ulf Blanke, Robert Rehner, Bernt Schiel * Max Planck Institute for Informatics ** DVS TU Darmstadt 15th IEEE International Symposium on Wearable Computers (ISWC), San Francisco, USA.
2010
Remember and Transfer what you have Learned - Recognizing Composite Activities based on Activity Spotting. Ulf Blanke, Bernt Schiele TU Darmstadt 14th IEEE International Symposium on Wearable Computers (ISWC), Seoul, Korea.
Characterizing Sleep Trends from Postures Marko Borazio*, Ulf Blanke, Kristof Van Laerhoven** * Max Planck Institute for Informatics ** ESS TU Darmstadt 14th IEEE International Symposium on Wearable Computers (ISWC), Seoul, Korea.
Towards Human Motion Capturing using Gyroscopeless Orientation Estimation. Ulf Blanke, Bernt Schiele TU Darmstadt 14th IEEE International Symposium on Wearable Computers (ISWC), Seoul, Korea.	Minute Madness Video Supplemental Video
Standing on the Shoulders of Other Researchers - A Position Statement. Ulf Blanke*, MIS, TU Darmstadt Diane Larlus, MIS, TU Darmstadt Kristof Van Laerhoven, ESS, TU Darmstadt Bernt Schiele, MIS, TU Darmstadt Pervasive Workshop "How to do good activity recognition research? Experimental methodologies, evaluation metrics, and reproducibility issues" Helsinki, Finland.
All for one or one for all? - Combining Heterogeneous Features for Activity Spotting. Ulf Blanke, Bernt Schiele TU Darmstadt Matthias Kreil, Paul Lukowicz ESL Uni Passau Thiemo Gruber, Bernard Sick CSL Uni Passau 7th IEEE PerCom Workshop on Context Modeling and Reasoning (CoMoRea).
2009
An Analysis of Sensor-Oriented vs. Model-Based Activity Recognition Andreas Zinnen SAP Research Ulf Blanke*, TU Darmstadt Bernt Schiele, TU Darmstadt 13th IEEE Intl. Symposium on Wearable Computing (ISWC) Linz, Austria.
Daily Routine Recognition through Activity Spotting Ulf Blanke*, TU Darmstadt Bernt Schiele, TU Darmstadt 4th Intl. Symposium on Location- and Context-Awareness (LoCA) Tokyo, Japan.
2008
Sensing Location in the Pocket Ulf Blanke, TU Darmstadt Bernt Schiele, TU Darmstadt 10th ACM Intl. Conference on Ubiquitous Computing (UbiComp) Seoul, South Korea. Poster*	Minute Madness Video
2007
Scalable Recognition of Daily Activities with Wearable Sensors Tam Huynh*, TU Darmstadt Ulf Blanke, TU Darmstadt Bernt Schiele, TU Darmstadt 3rd Int. Symposium on Location- and Context-Awareness (LoCA) Oberpfaffenhofen, Germany.
* Presenter. For further information (,e.g., datasets etc.): http://www.mis.tu-darmstadt.de Other publications & talks

Visualization as Tool for Data Understanding Ulf Blanke Summer School “Visualizations of Space: Concepts-Methods-Media” 2010, Ebernburg, Germany Within interdisciplinary research project “Visualization – Information – Learning/Perception” – together with Anna Lisa Martin (TU Darmstadt) and Markus Stross (TU Darmstadt)
Understanding the process of technology researchand its impact on the real world Ulf Blanke, MIS, TU Darmstadt Suzana Alpsancar, TU Darmstadt Young Philosophy Second (under)-graduate conference, 2010, Darmstadt, Germany
Unsupervised Activity Recognition using Wearable Sensors Diploma thesis Implementation of HMM-based EM clustering algorithm to recognize daily activities, e.g., ironing, eating and scenes, such as at work, housekeeping. B. Clarkson & A. Pentland. Unsupervised clustering of ambulatory audio and video. In icassp, 1999. Brian Clarkson, Kenji Mase & Alex Pentland. Recognizing User Context via Wearable Sensors. In ISWC, 2000.

Research

Dissertation

Recognizing Complex Human Activity Based on Activity Spotting

Peer reviewed publications

Crowdsourced Pedestrian Map Construction for Short-Term City-Scale Events

Inferring Travel Purpose from Crowd-Augmented Human

Participatory Bluetooth Scans Serving as Urban Crowd Probes

How to Save Human Lives with Complexity Science

Capturing crowd dynamics at large scale events using participatory GPS-localization

GaitAssist: A Daily-Life Support and Training System for Parkinson’s Disease Patients with Freezing of Gait

(Demo) GaitAssist: A Wearable Assistant for Gait Training and Rehabilitation in Parkinson's Disease

The Telepathic Phone: Frictionless Activity Recognition from WiFi-RSSI

A Tutorial on Human Activity Recognition Using Body-worn Inertial Sensors

Mobile Computing, Applications, and Services

Human Activity Recognition Using Social Media Data

Towards Scalable Activity Recognition: Adapting Zero-Effort Crowdsourced Acoustic Models

Low-Power Ambient Sensing in Smartphones for Continuous Semantic Localization

Combining Crowd-Generated Media and Personal Data: Semi-Supervised Learning for Context Recognition

Prior Knowledge of Human Activities from Crowd-Generated Social Data

Engineers Meet Clinicians: Augmenting Parkinson's Disease Patients to Gather Information for Gait Rehabilitation

User Independent, Multi-Modal Spotting of Subtle Arm Actions with Minimal Training Data.

Ah! I felt like doing something else...

I spy with my little eye: Learning Optimal Filters for Cross-Modal Stereo under Projected Patterns

Improving the Kinect by Cross-Modal Stereo

South by South-East Or Sitting at the Desk. Can Orientation be a place?

Remember and Transfer what you have Learned - Recognizing Composite Activities based on Activity Spotting.

Characterizing Sleep Trends from Postures

Towards Human Motion Capturing using Gyroscopeless Orientation Estimation.

Standing on the Shoulders of Other Researchers - A Position Statement.

All for one or one for all? - Combining Heterogeneous Features for Activity Spotting.

An Analysis of Sensor-Oriented vs. Model-Based Activity Recognition

Daily Routine Recognition through Activity Spotting

Sensing Location in the Pocket

Scalable Recognition of Daily Activities with Wearable Sensors

Other publications & talks

Visualization as Tool for Data Understanding

Understanding the process of technology research and its impact on the real world

Unsupervised Activity Recognition using Wearable Sensors

Recognizing Complex Human Activity
Based on Activity Spotting

Understanding the process of technology researchand its impact on the real world