F. Mokhtarian and S. Abbasi (UK)
Shape Description, Fusion, Robust Recognition,Curvature Scale Space, MPEG-7 standard
This paper addresses the issue of automatic selection of the best and the optimum number of views for each ob ject in multi-view 3-D object representation and recogni tion. The object boundary of each view is considered as a 2-D shape and is represented by the maxima of its Cur vature Scale Space (CSS) image contours. The CSS rep resentation has been selected for MPEG-7 standardization as a contour shape descriptor. After representing each ob ject by an optimal number of views, an unknown object is recognised by a single image taken from an arbitrary view point using the same shape representation. The paper also presents a novel method for fusion of results from multi ple shape descriptors. The utilisation of this method for multi-view 3-D object representation and recognition has been explored. The object boundary of each view is repre sented effectively using the CSS technique, Moment invari ants, and Fourier Descriptors. The results obtained from the fusion method are superior to the results obtained from any single technique. The method has been tested on a col lection of 3-D objects consisting of aircrafts of different shapes. An analysis of the performance of the system has been given in this paper as the number of views varies.
Important Links:
Go Back
