Context-adaptive navigation of 3D model collections

Abstract

When reasoning about similarity in a collection of objects with heterogeneous qualities, there are several aspects of interest that can be followed to explore the collection. Indeed, the notion of similarity among 3D models is not only grounded on the geometric shape but also, for instance, on the style, material, color, decorations, common parts. These are all important factors that concur to the concept of similarity. Search engines for visual content are expected to address similarity assessment in collections, providing a higher degree of flexibility with respect to the traditional 3D object retrieval operations. In this work, we describe the design and functioning of a search engine working on multiple factors and discuss the results on a number of collections, which challenge existing 3D object retrieval engines.

A New Partial 3D Object Indexing and Retrieval Approach Combining 2D slices and Apriori Algorithm

“…They adopted an unsupervised approach for learning shape descriptors using sparse autoencoders. Biasotti et al [7] presented an approach, a search engine model for dataset exploration, which is based on multiple similarity criteria between models in its search. The combination of similarity criteria they are proposing is a user-driven navigation and similarity assessment.…”

Foreword to the Special Section on 3D Object Retrieval (3DOR2018)

“…Biasotti et al [2] address the important problem of searching large 3D shape collections in the context of faceted queries and automatic search relaxation in content-based shape retrieval. They show how search results can be improved by considering richer, more complex, ten-dimensional descriptors.…”