Abstract
The LBP scale space serves as a shaperepresentation, which thus allows not only for shapeclassification but also for an (approximate) recon-struction of the original shape. In this paper possibleLBP scale space origins within a shape are evalu-ated. The influence of the LBP scale space centeron the reconstruction quality is studied by comput-ing the LBP scale space for every position inside ashape and by comparing the according reconstruc-tion errors. For shape classification, the LBP scalespace is further evaluated on the MPEG-7 dataset.
Reference
Janusch, I., & Kropatsch, W. (2017). LBP Scale Space Origins for Shape Classification. In Proceedings of the 22nd Computer Vision Winter Workshop 2017 (p. 9). http://hdl.handle.net/20.500.12708/57084