TNT logo LUH TNT
Dr.-Ing. Florian Baumann
Florian Baumann left the Institut für Informationsverabeitung.
Publications and research activities from the time after the departure are not listed here.

Florian Baumann studied Electrical Engineering with a focus on computer engineering at the Leibniz University of Hannover. He received his Dipl.-Ing. in March of 2012. Since May of 2012 Florian Baumann is working toward a doctor's degree (Dr.-Ing.) in the group of Prof. Rosenhahn at the eNIFE.
His research interests are image processing, automated image interpretation and computer vision, e.g.:

Show selected publications only
  • Alfred Oliver Effenberg, Gerd Schmitz, Florian Baumann, Bodo Rosenhahn, Kroeger Daniela
    SoundScript–Supporting the Acquisition of Character Writing by Multisensory Integration
    Open Psychology Journal, Vol. 8, No. 3, pp. 230-237, January 2016
  • Christoph Reinders, Florian Baumann, Björn Scheuermann, Arne Ehlers, Nicole Mühlpforte, Alfred O. Effenberg, Bodo Rosenhahn
    On-The-Fly Handwriting Recognition using a High-Level Representation
    The 16th International Conference on Computer Analysis of Images and Patterns (CAIP), Lecture Notes in Computer Science (LNCS), Valetta, Malta, September 2015
  • Florian Baumann, Jinghui Chen, Karsten Vogt, Bodo Rosenhahn
    Improved Threshold Selection by using Calibrated Probabilities for Random Forest Classifiers
    12th Conference on Computer and Robot Vision (CRV), Halifax, Nova Scotia, Canada, June 2015
  • Florian Baumann, Karsten Vogt, Arne Ehlers, Bodo Rosenhahn
    Probabilistic Nodes for Modelling Classification Uncertainty for Random Forest
    14th IAPR International Conference on Machine Vision Applications (MVA), Tokio, Japan, May 2015
  • Florian Baumann, Jie Liao, Arne Ehlers, Bodo Rosenhahn
    Recognizing Human Actions using novel Space-time Volume Binary Patterns
    Neurocomputing Journal (to appear), April 2015
  • Florian Baumann, Liu Wei, Arne Ehlers, Bodo Rosenhahn
    Sequential Boosting for Learning a Random Forest Classifier
    IEEE Winter Conference on Applications of Computer Vision (WACV), Waikoloa Beach, HI, USA, January 2015
  • Florian Baumann, Irina Schulz, Bodo Rosenhahn
    Multi-Sensor Acceleration-based Action Recognition
    International Conference on Image Analysis and Recognition (ICIAR), October 2014
  • Florian Baumann, Jie Liao, Arne Ehlers, Bodo Rosenhahn
    Computation Strategies for Volume Local Binary Patterns applied to Action Recognition
    11th IEEE International Conference on Advanced Video and Signal-Based Surveillance , Seoul, Korea, August 2014
  • Florian Baumann, Jie Liao, Arne Ehlers, Bodo Rosenhahn
    Motion Binary Patterns for Action Recognition
    3rd International Conference on Pattern Recognition Applications and Methods, France, Angers Loire Valley, March 2014
  • Florian Baumann, Li Fangda, Arne Ehlers, Rosenhahn Bodo
    Thresholding a Random Forest Classifier
    Advances in Visual Computing - 10th International Symposium, Springer, Las Vegas, NV, USA, 2014, edited by George Bebis et al.
  • Arne Ehlers, Florian Baumann, Bodo Rosenhahn
    Boosted Fractal Integral Paths for Object Detection
    Advances in Visual Computing - 10th International Symposium, Springer, Las Vegas, NV, USA, 2014, edited by George Bebis et al.
  • Arne Ehlers, Björn Scheuermann, Florian Baumann, Bodo Rosenhahn
    Cleaning Up Multiple Detections Caused by Sliding Window Based Object Detectors
    18th Iberoamerican Congress on Pattern Recognition (CIARP), Havanna, Cuba, November 2013
  • Florian Baumann
    Action Recognition with HOG-OF Features
    35th German Conference on Pattern Recognition (YRF at GCPR), Saarbrücken, Germany, September 2013
  • Florian Baumann, Arne Ehlers, Karsten Vogt, Bodo Rosenhahn
    Cascaded Random Forest for Fast Object Detection
    18th Scandinavian Conference on Image Analysis (SCIA), Espoo, Finland, June 2013
  • Arne Ehlers, Florian Baumann, Bodo Rosenhahn
    Exploiting Object Characteristics using Custom Features for Boosting-based Classification
    18th Scandinavian Conference on Image Analysis (SCIA), Espoo, Finland, June 2013
  • Arne Ehlers, Florian Baumann, Ralf Spindler, Birgit Glasmacher, Bodo Rosenhahn
    PCA Enhanced Training Data for Adaboost
    Computer Analysis of Images and Patterns - 14th International Conference, CAIP 2011, Seville, Spain, August 29-31, 2011, Proceedings, Part I, Springer, Vol. 6854, pp. 410-419, August 2011
  • Björn Scheuermann and Arne Ehlers and Hamon Riazy and Florian Baumann and Bodo Rosenhahn
    Ego-Motion Compensated Face Detection on a Mobile Device
    IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops; The Seventh IEEE Workshop on Embedded Computer Vision (ECVW2011), June 2011
  • Florian Baumann, Katharina Ernst, Arne Ehlers, Bodo Rosenhahn
    Symmetry Enhanced Adaboost
    Advances in Visual Computing - 6th International Symposium, Springer, Vol. 6453, pp. 286--295, Las Vegas, NV, USA, November 2010, edited by George Bebis et al.
Other activities

Auszeichnungen:

Supervised theses:

  • Evaluation of different LBP Computation Strategies for Action Recognition [abgeschlossen]
  • Kinect-based Action Recognition [abgeschlossen]
  • Evaluation of HOG3D and Random Forest for Action Recognition [abgeschlossen]
  • Acceleration-based Action Recognition [abgeschlossen]
  • Feature Evaluation for Trafficsign Recognition [abgeschlossen]
  • Eye Movement Detection on Mobile Devices [abgeschlossen]
  • An Augmented Reality Game for Mobile Devices [abgeschlossen]
  • Object Detection and Classification in Thermal Images [abgeschlossen]
  • Slow-Feature Analysis for Action Recognition [abgeschlossen]
  • Evaluation of Machine Learning Algorithms for Handwriting Recognition [in Bearbeitung]
  • Soccer Ball Detection for Robo-Cup [abgeschlossen]
  • A Spacetime Feature for Action Recognition [abgeschlossen]
  • User-interacted Multi-class Object Detection on a Mobile Device [abgeschlossen]
  • Action Bank for Human Action Recognition [in Bearbeitung]

Open theses:

  • XSens accelerations sensors
  • Eye tracking on mobile devices (iOS and Android)
  • Algorithmic modifications (AdaBoost, Random Forest)
  • Kinect Action Recognition
  • ToF Action Recognition (with Time of Flight camera)
  • Modify existing framework for Action Recognition with Volume Local Binary Patterns
  • Implement interactive games on mobile devices (iOS and Android) using object detection frameworks

Contact me for further information and bachelor or master thesis proposals.



Projects: Object Detection and Classification in Thermal Images:





Projects: Object Detection with Aerial Drone:




ARF-OktoXL by Mikrokopter


Projects: Action Recognition with Inertial Sensors:




Acceleration-based with inertial sensor by http://www.xsens.com

Projects: Object Detection and Recognition:






Projects: Action Recognition:







Projects: Interactive Games on Mobile Devices:





Open positions and theses:


GPU basierte Objektdetektion
Sonstige Programmiertätigkeiten im Bereich Objektdetektion

Open Bachelor- and Mastertheses:


Android Software Entwickler
Action Recognition mit 2-3 Referenzverfahren
Multiple Objekterkennung mit Random Forest




Source code:


This code is provided "as is" without any implied warranty for non-commercial use only. Permission to modify and distribute this code is granted. If you use this software or any modifications, please cite the corresponding paper.

For the source code of the ICPRAM'14 paper please write a mail.


Annotations:


Weizman annotations for the ICPRAM'14 paper: https://db.tt/7rTr96i7


Poster presentations:


Download GCPR 2012 poster


Oral presentations:

  • ISVC'2010
  • SCIA'2013
  • ICPRAM'2014
  • AVSS'2014
  • ICIAR'2014
  • ISVC'2014
  • WACV'2015




Founder of the spin-off HearingCampus.com:


Geschäftsidee ist eine App, mit der Jedermann zuhause sein eigenes Gehör testen kann. Die Möglichkeiten gehen dabei über einen normalen Hörtest hinaus: so kann mit unserer Lösung auch das Verständnis von Silben und Zahlen untersucht werden. Außerdem kann direkt aus der App heraus Kontakt zu einem Hörgeräteakustiker in der Nähe aufgebaut werden, wenn eine Hörminderung festgestellt werden sollte. Der Akustiker kann dann das zuvor mit der App erstellte Hörprofil auf Wunsch des Kunden auslesen und als Grundlage für die Einstellung von Hörgeräten verwenden. So sparen beide Seiten Zeit und es kann eine optimale Anpassung auf die Hörbedürfnisse des Kunden vorgenommen werden - bestehende Hörgeräteträger können die App nutzen, um ihr Hörvermögen laufend zu untersuchen und wenn nötig die Einstellungen ändern lassen. Dies schafft effektivere Arbeitsprozesse für den Akustiker und erhöht gleichzeitig die Zufriedenheit der Kunden. Wir agieren dabei als unabhängiges Bindeglied zwischen beiden Gruppen.