In many application domains such as Computer Aided Design, Architecture, Visualization, Entertainment and so forth, 3D models are being created, leading to 3D repositories of increasing size. Also, improved acquisition and reconstruction methods contribute to an increase in the number of available 3D model data. As 3D repositories continue to increase in size, methods for content-based retrieval in 3D model data become increasingly important. In this project, we research methods for 3D model retrieval based on global shape descriptors. Recent work of ours explored the application of 2D image analysis methods in the 3D domain, including descriptors based on the Hough Transform. We also research methods for combining sets of complementary 3D descriptors for improved retrieval results.
Contact: Tobias Schreck, Maximilian Scherer
Publications
STELA: Sketch-Based 3D Model Retrieval using a Structure-Based Local Approach
Saavedra, J. M.; Bustos, B.; Scherer, M.; Schreck, T.
In: ACM International Conference on Mulimedia Retrieval , to appear, 2011
Sketch-based 3D model retrieval using diffusion tensor fields of suggestive contours
Yoon, S. M.; Scherer, M.; Schreck, T.; Kuijper, A.
In: Proceedings of the international conference on Multimedia (ACM MM), pp. 193-200, 2010
Histograms of Oriented Gradients for 3D Object Retrieval
Scherer, M.; Walter, M.; Schreck, T.
WSCG 2010, 18th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision. Full Papers Proceeding, pp. 41-48, 2010
