parlab 
research> activities |
Despite the recent developments in three-dimensional (3D) computer vision, it becomes apparent that high performance of automatic vision in some aspects contrasts with its poor performance in other aspects. Combining automatic and human vision appears as a challenging way towards improved complex 3D vision. The general aim of this project is to investigate the combined use of human and computer 3D vision for improved performance. The goal is to solve positioning and shape recognition tasks for complex 3D objects.
The methodology is an experimental one. Human interaction will be mostly used for its effectiveness to solve global search problems and will take place in the context of a virtual or augmented reality environment. Computer vision will dominate in the shape matching problem. Advanced approaches based on geometric matching that suit arbitrary complex objects will be applied.
We propose to develop two experimental platforms. Application fields for the developed methods are robotics in presence of CAD-CAM, telemanipulation together with virtual reality, recognition of complex shapes like footprints, biological structures, micro-structures.
This project concerns the vision of free-form three dimensional (3D) objects. The general aim was to combine human interpretation and advanced computer capabilities to solve the task. A system was investigated and developed that takes the best of both worlds. It permits to solve various tasks that cannot be solved in the framework of classical vision.
The developed system allows to scan 3D objects in a very unconstrained way and to perform various kinds of vision tasks: object recognition, object modeling and scene reconstruction. The developed innovative methods permit fast and accurate geometric matching and allow working with colored and textured objects. Building virtual objects by digitizing and modeling physical objects was demonstrated on several objects and scenes. The methods and system are operational in various application fields like museography, multimedia, reverse engineering, virtual world construction, quality inspection
Fig. 1: Geometric shape matching permits to recognize free-form objects
C. Schütz, T. Jost and H. Hügli, "Multi-feature matching algorithm for free-form 3D surface registration", Proc. ICPR, Australia, pp 982-984, Aug. 1998
C. Schütz, " Geometric Point Matching of free-form 3D objects", PhD thesis, Université de Neuchâtel, July 1998
A. Sulzmann, Ch. Schütz, H. Hügli, J. Jacot, "Augmented reality as an interactive tool for microscopic imaging, measurement and model based verification of simulated parts", Int. Conf. Modeling and Simulation of Microsystems, MSM '98, April 4-6, 1998, Santa Clara
C. Schütz, T. Jost and H. Hügli, "Semi-Automatic 3D Object Digitizing System Using Range Images", Proc. ACCV'98, Hong-Kong, pp. 490-497, Jan 1998
Christian Schütz, Timothée Jost, Heinz Hügli, "Free-form 3D object reconstruction from range images", Proc. Int. Conf. Virtual Systems and MultiMedia, VSMM'97, Geneva, IEEE Computer Society, pp. 69-70, Aug. 1997
Christian Schütz, Heinz Hügli, "Augmented reality using range images", SPIE Photonics West, The Engineering Reality of Virtual Reality, San Jose, Proc. SPIE, Vol. 3012- IV, pp. 472-478, Feb. 1997
Heinz Hügli, Christian Schütz, "Geometric matching of 3D objects: assessing the range of successful configurations", Proc. International Conference of Recent Advances in 3-D Digital Imaging and Modeling, Ottawa, Ontario, Canada, pp 101-106, May 12-15, 1997
Christian Schutz, Emerico Natonek, Charles Baur, Heinz Hugli, "3D Vision in a Virtual Reality Robotics Environment", SPIE Photonics East, Telemanipulator and Telepresence Technologies III, Boston, Proc. SPIE, Vol. 3012, pp. 138-146, Nov. 1996
Heinz Hügli, Christian Schutz, "How well performs free-form 3D object recognition from range images?", Proc. SPIE Photonics East, Vol 2904-09, pp. 66-74, Boston, Nov.1996
Heinz Hügli, Christian Schütz, "Computer vision of free-form 3D objects by geometric matching", Mecatronics, Europe-Asia congress on Mechatronics, Besançon, Oct. 1-3 1996
Heinz Hügli, "Knowledge-based 3D vision systems", Proc. Closing conference Priority program informatics research, EPF Lausanne, March 19 & 20, 1996
Heinz Hügli, Christian Schütz, "Change detection in range imaging for 3D scene segmentation", SPIE Europto, Vol 2786, 10-14 June 1996, Besançon
Heinz Hügli, Christian Schütz, Dimitrios Semitekos, "Geometric matching for free-form 3D object recognition", Proc. ACCV'95, Vol III, pp 819-823, 5-8 December 1995
Christian Schutz, Heinz Hugli, "Towards the recognition of 3D free-form objects", SPIE Int. Symposium on Intelligent Robots and Advanced Manufacturing, Intelligent Robots and Computer Vision XIV, Philadelphia, Oct. 1995
|
hu / |