Title :
A passive full body scanner using shape from silhouettes
Author_Institution :
Inst. fur Theoretische Nachrichtentechnik und Informationsverarbeitung, Hannover, Germany
Abstract :
Describes a camera-based approach to fully automatically create 3D models of persons. A setup of sixteen digital cameras is used to capture the images of the person to be scanned. Using a monochromatic background and the shape-from-silhouette approach a 3D model is created automatically. Using the original images the models are realistically textured. In future 3D telepresence applications or multi-player games these models have to be animated using an internal skeleton structure. A principal component analysis is utilised to extract the skeleton automatically from the models. Compared to other full body scanning approaches no mechanically moving components are used which makes the system especially suitable for low cost end user applications
Keywords :
feature extraction; image segmentation; image sensors; image texture; principal component analysis; rendering (computer graphics); 3D models; 3D telepresence; camera-based approach; internal skeleton structure; monochromatic background; multi-player games; passive full body scanner; shape-from-silhouette approach; Animation; Calibration; Digital cameras; Layout; Shape; Skeleton; Streaming media; Video sharing; Videoconference; Virtual environment;
Conference_Titel :
Pattern Recognition, 2000. Proceedings. 15th International Conference on
Conference_Location :
Barcelona
Print_ISBN :
0-7695-0750-6
DOI :
10.1109/ICPR.2000.905495