3D Shape Indexing and Retrieval Using Characteristics level images
In this paper, we propose an improved version of the descriptor that we proposed before. The descriptor is based on a set of binary images extracted from the 3D model called level images noted LI. The set LI is often bulky, why we introduced the X-means technique to reduce its size instead of K-means used in the old version. A 2D binary image descriptor was introduced to extract the vectors descriptors of the 3D model. For a comparative study of two versions of the descriptor, we used the National Taiwan University (NTU) database of 3D object.
Keywords: 3D shape retrieval, shape descriptor, Zernike moments, characteristics level images, NTU database, X-means.
Download Full-Text
ABOUT THE AUTHORS
Abdelghni Lakehal
GRMS2I, Sidi Mohamed Ben AbdEllah University Fez , B.P 1796, Morocco
Omar El Beqqali
GRMS2I, Sidi Mohamed Ben AbdEllah University Fez , B.P 1796, Morocco
Abdelghni Lakehal
GRMS2I, Sidi Mohamed Ben AbdEllah University Fez , B.P 1796, Morocco
Omar El Beqqali
GRMS2I, Sidi Mohamed Ben AbdEllah University Fez , B.P 1796, Morocco