The definition of reliable shape descriptors is an essential topic for 3D object retrieval. In general, two main approaches are considered: global, and local. Global approaches are effective in describing the whole object, while local ones are more suitable to characterize small parts of the shape. Recently some strategies to combine these two approaches have been proposed which are mainly concentrated to the so-called bag of words paradigm. With this paper we address this problem and propose an alternative strategy that goes beyond the bag of word approach. In particular, a sparse coding technique is exploited for the 3Ddomain: a set of local shape descriptors are collected fromthe shape, and then a dictionary is trained as generativemodel. In this fashion the dictionary is used as global shapedescriptor for shape retrieval purposes. Several experimentsare performed on standard databases in order to evaluate theproposed method in challenging situations like the case of‘SHREC 2011: robustness benchmark’ where strong shapetransformations are included, and the case of ‘SHREC 2007:partial matching track’ where composite models are consideredin the query phase. A drastic improvement of theproposed method is observed by showing that sparse codingapproach is particularly suitable for local-to-global descriptionand outperforms other approaches such as the bag ofwords.
A sparse coding approach for local-to-global 3D shape description
CASTELLANI, Umberto
2014-01-01
Abstract
The definition of reliable shape descriptors is an essential topic for 3D object retrieval. In general, two main approaches are considered: global, and local. Global approaches are effective in describing the whole object, while local ones are more suitable to characterize small parts of the shape. Recently some strategies to combine these two approaches have been proposed which are mainly concentrated to the so-called bag of words paradigm. With this paper we address this problem and propose an alternative strategy that goes beyond the bag of word approach. In particular, a sparse coding technique is exploited for the 3Ddomain: a set of local shape descriptors are collected fromthe shape, and then a dictionary is trained as generativemodel. In this fashion the dictionary is used as global shapedescriptor for shape retrieval purposes. Several experimentsare performed on standard databases in order to evaluate theproposed method in challenging situations like the case of‘SHREC 2011: robustness benchmark’ where strong shapetransformations are included, and the case of ‘SHREC 2007:partial matching track’ where composite models are consideredin the query phase. A drastic improvement of theproposed method is observed by showing that sparse codingapproach is particularly suitable for local-to-global descriptionand outperforms other approaches such as the bag ofwords.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
