Научная визуализация, 2020, том 12, номер 2, страницы 110 - 126, DOI: 10.26583/sv.12.2.09
A New Partial 3D Object Indexing and Retrieval Approach Combining 2D slices and Apriori Algorithm
Авторы: I.O. Taybi1,A, T. Gadi2,A, R. Alaoui3,B,C
A Laboratory of Informatics, Imaging, and Modeling of Complex Systems (LIIMSC) Faculty of Sciences and Techniques, Hassan 1st University, Settat, Morocco
B LRIT Laboratory, Faculty of Sciences, Mohammed V University, Rabat Morocco
C LASTIMI Laboratory, Higher School of Technology - Sale, Mohammed V University, Rabat Morocco
1 ORCID: 0000-0001-6782-9807, ilyass.ouazzani@gmail.com
2 ORCID: 0000-0002-2174-5816, gtaoufiq@yahoo.fr
3 ORCID: 0000-0001-9948-329X, alaoui.rach@gmail.com
Аннотация
This paper examines the issue of 3D object indexing and retrieval and tries to solve this problem using partial indexing approach. The hypothesis in this context is that similar 3D objects will be composed of similar 2D slices. The proposed partial 3D object indexing and retrieval method is applicable on both complete and incomplete 3D objects, which is based on a similarity measuring between 2D slices of 3D objects. The main idea behind our approach is to extract an initial set of 2D slices corresponding to determined axes, and then use the Apriori algorithm to select the most representative ones, transforming the issue of shape-matching between 3D objects into evaluating the similarities between their 2D slices. Experiments on the Princeton Shape Benchmark (PSB) indicate that our approach outperforms evaluated retrieval approaches.
Ключевые слова: 3D object indexing, 3D object retrieval, Cluster validity index, Data mining, Apriori Algorithm, Association rules, Partial similarity.
