StefanHinterstoisser

Chair for Computer Aided Medical Procedures & Augmented Reality
Lehrstuhl für Informatikanwendungen in der Medizin & Augmented Reality

Contact

Dr. Dipl.-Inf. Univ. Stefan Hinterstoißer
email
phone +49 (89) 289-19400
fax +49 (89) 289-17059
room 03.13.057
address Institut für Informatik I-16
Technische Universität München
Boltzmannstr. 3
85748 Garching b. München
Germany
skype My status
HinterstoisserStefanicon2.JPG

Research

  • My Ph.D. Advisors were Prof. Nassir Navab, Dr. Vincent Lepetit, Dr. Slobodan Ilic and Dr. Selim Benhimane. I defended my thesis on 14.03.2012. My committee was Prof. Bernt Schiele, Prof. Kurt Konolige, Prof. Michael Beetz and Prof. Nassir Navab.
  • Fields of Interest
    • 2D / 3D Object Detection
    • RGBD Devices
    • Pose Estimatin
    • 3D Model Creation (In-hand, Markerboard dependent)
    • Real-Time

Miscellaneous

Reviewer for: PAMI, CVPR, ECCV, ICPR, ISMAR, ICCV, BMVC, T-RO, VCG

Supported by the BMBF project AVILUSplus (01IM08002) until 28.02.2011. Supported by WillowGarage? from 01.03.2011 until 30.09.2012

Work2.gif

Teaching Assistance

Student Projects

Available

There are currently no projects available.

Finished

  • Martin Saelzle: In-hand Scanning of Texture-less 3D Objects (Master Thesis)
  • Stefan Holzer: Distance Transform Templates for Object Detection and Pose Estimation (Master Thesis)
  • Nicolas Alt: Optimal Template Selection for Visual Tracking (Master Thesis)
  • Ya Chen: Evaluation of Real-Time Object Detection and Tracking Method for Defining an Optimal Performance Algorithm (Master Thesis)

News

  • [04.10.2012] Our partners from TOYOTA are using Linemod for their robots. Check out this video!
  • [30.09.2012] Our ACCV dataset is out!
  • [21.09.2012] Our ACCV paper is accepted! We will demonstrate it at ECCV in Firence!
  • [05.06.2012] We got our ECCV Demo accepted! See you there!
  • [03.04.2012] LINE-MOD code is finally available. Just check out the latest opencv version via svn!
  • [21.09.2011] We got our TPAMI paper accepted!
  • [27.06.2011] We got our ICCV paper about our new method LINEMOD accepted as Oral Presentation. It is about combining 2D color gradients with 3D surface normals (from the Kinect depth map) for robust detection of texture-less Objects under heavily cluttered background. As usual we will release source code (however that might still take some time - stay tuned!!!). This approach is much better than DOT (much more robust, much less false positives, runtime independent of template size etc.)! See the paper and the videos!
  • [16.02.2011] Wonwoo ported our DOT software to Mac OS. Please refer to his blog for further information.
  • [06.12.2010] Steve made our DOT software running under Linux. Please refer to his blog for further information.
  • [19.08.2010] Our paper on "Learning Real-Time Perspective Patch Rectification" has been accepted for publication in International Journal of Computer Vision (IJCV).

Software

LINE-MOD code is finally available (however in beta release only). Please note that the code does not have the same gradient online learning method as we used in ICCV11! Some other minor features are also different (e.g. thresholding). The current gradient method selects gradients on the silhouette only and is made for artificial 3D objects! It is not suited for online learning or for versions using color information only! The version we used at ICCV11 selects strong gradients anywhere on the object. Unfortunately, the opencv code will not allow you to reproduce the results of my papers (the results will be way worse!)! Apart from it - we changed the licensing for the current software packages from GPL to LGPL! We have not updated the source files correspondingly but will do it somewhen!

All software (LGPL licensed) can be found HERE.

Database

The extensive ACCV database is public (over 18000 real images with 15 different objects and ground truth pose)! Here you can find the databases for the ape, benchvise, bowl, can, cat, cup, driller, duck, glue, holepuncher, iron, lamp, phone, cam and eggbox. Each dataset contains the 3D model saved as a point cloud (format: #_of_voxels size_of_voxel_in_cm x1_in_cm y1_in_cm z1_in_cm normal_x1 normal_y1 normal_z1 color_x1_normalized_to_1 color_y1_normalized_to_1 color_z1_normalized_to_1 ...) and a file that contains called distance.txt with the maximum diameter of the object (in cm). For some datasets we also provide a nice mesh model in the ply format (in mm - with better normals). The original mesh is contained in OLDmesh.ply. For most datasets we registered this OLDmesh.ply to the point cloud with the transformation stored in transform.dat (first number is not important, then each first number of a line is obsolete - for the rest: the transformation matrix [R|T] is stored rowwise (in m)). The registered mesh is stored in mesh.ply. In the folder data you can find the color images, the aligned depth images and the ground truth rotation and translation (in cm). In order to read the depth images you can use this function. The internal camera matrix parameters for the kinect are: fx=572.41140, px=325.26110, fy=573.57043; py=242.04899; Color image and depth image are already aligned by the internal alignement procedure of Kinect.

Publications

Journal Articles

2012
S. Hinterstoisser, C. Cagniart, S. Ilic, P. Sturm, N. Navab, P. Fua, V. Lepetit
Gradient Response Maps for Real-Time Detection of Texture-Less Objects
IEEE Transactions on Pattern Analysis and Maschine Intelligence (TPAMI). (bib)
2011
S. Hinterstoisser, V. Lepetit, S. Benhimane, P. Fua, N. Navab
Learning Real-Time Perspective Patch Rectification
International Journal of Computer Vision (IJCV), Springer Verlag, The final publication is available at www.springerlink.com (bib)

International Conferences

2012
S. Hinterstoisser, V. Lepetit, S. Ilic, S. Holzer, G. Bradski, K. Konolige, N. Navab
Model Based Training, Detection and Pose Estimation of Texture-Less 3D Objects in Heavily Cluttered Scenes
Asian Conference on Computer Vision (ACCV), Korea, Daejeon, November 2012 (bib)
2011
S. Hinterstoisser, S. Holzer, C. Cagniart, S. Ilic, K. Konolige, N. Navab, V. Lepetit
Multimodal Templates for Real-Time Detection of Texture-less Objects in Heavily Cluttered Scenes (Oral)
IEEE International Conference on Computer Vision (ICCV), Barcelona, Spain, November 2011. (bib)
2010
N. Alt, S. Hinterstoisser, N. Navab
Rapid Selection of Reliable Templates for Visual Tracking
IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), San Francisco, California (USA), June 2010. (bib)
S. Hinterstoisser, V. Lepetit, S. Ilic, P. Fua, N. Navab
Dominant Orientation Templates for Real-Time Detection of Texture-Less Objects
IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), San Francisco, California (USA), June 2010. (bib)
2009
S. Hinterstoisser, O. Kutter, N. Navab, P. Fua, V. Lepetit
Real-Time Learning of Accurate Patch Rectification (Oral Presentation)
IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), Miami, Florida (USA), June 2009. (bib)
S. Holzer, S. Hinterstoisser, S. Ilic, N. Navab
Distance Transform Templates for Object Detection and Pose Estimation
IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), Miami, Florida (USA), June 2009. (bib)
2008
S. Hinterstoisser, S. Benhimane, V. Lepetit, P. Fua, N. Navab
Simultaneous Recognition and Homography Extraction of Local Patches with a Simple Linear Classifier
British Machine Vision Conference (BMVC), Leeds (UK), September 1-4, 2008 (bib)
S. Hinterstoisser, S. Benhimane, N. Navab, P. Fua, V. Lepetit
Online Learning of Patch Perspective Rectification for Efficient Object Detection
IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Anchorage, Alaska (USA), June 2008. (bib)
2007
P. Georgel, P. Schroeder, S. Benhimane, S. Hinterstoisser, M. Appel, N. Navab
An Industrial Augmented Reality Solution For Discrepancy Check
Proceedings of the 6th International Symposium on Mixed and Augmented Reality (ISMAR), Nara, Japan, Nov. 2007 (bib)
S. Hinterstoisser, S. Benhimane, N. Navab
N3M: Natural 3D Markers for Real-Time Object Detection and Pose Estimation
IEEE International Conference on Computer Vision, Rio de Janeiro, Brazil, October 14-20, 2007 (bib)

Awards

Pose Estimation / Matching

Model-based Detection and Pose Estimation (ACCV12 - realistic 3D Models are augmented on the image)

img2.jpg img3.jpg

LINE-MOD (ICCV11)

linemod1.jpg linemod2.jpg

Dominant Orientation Templates (CVPR10)

dot1.jpg dot2.jpg

Distance Transform Templates (CVPR09)

dtt1.jpg dtt2.jpg

Gepard (CVPR09)

gepard1.png gepard2.png

Panter (BMVC08)

panter1.png panter2.png

Leopar (CVPR08)

leopar1.jpg leopar2.jpg

N3Ms (ICCV07)

N3M.jpg N3M1.jpg



UsersForm
Title: Dr.
Circumference of your head (in cm):  
Firstname: Stefan
Middlename:  
Lastname: Hinterstoisser
Picture: HinterstoisserStefanicon1.jpg
Birthday:  
Nationality: Germany
Languages:  
Groups: Computer Vision
Expertise: Computer Vision
Position: Scientific Staff
Status: Alumni
Emailbefore: hinterst
Emailafter: in.tum.de
Room: MI 03.13.042
Telephone: +49 89 289 19400
Alumniactivity:  
Defensedate:  
Thesistitle:  
Alumnihomepage:  
Personalvideo01:  
Personalvideotext01:  
Personalvideopreview01:  
Personalvideo02:  
Personalvideotext02:  
Personalvideopreview02:  


Edit | Attach | Refresh | Diffs | More | Revision r1.112 - 29 Oct 2013 - 11:49 - StefanHinterstoisser

Lehrstuhl für Computer Aided Medical Procedures & Augmented Reality    rss.gif