Content-based 3D image retrieval using point cloud library a novel approach for the retrieval of 3D images

Abstract

Retrieval by content of 3D images has become an area of utmost interest at present. Unlike earlier times, 3D imaging is more affordable owing to the reduction in the software and hardware costs. Hence there has been an increasing degree of technological advancements in this field trying to meet the increasing demand. One such area of interest is content-based image retrieval i.e. obtaining similar 3D images on providing an input 3D image. In content-based image retrieval, the features like texture and shape of the input image are considered for the retrieval process. In this paper, a method for content-based 3D image retrieval using point cloud library is put forth. Point Cloud Library (PCL) is an open project for 2D/3D images and point cloud processing. Here different stages of retrieving the 3D image are by concentrating on 4 major steps i.e. key-point detection, feature extraction, comparison and visualization. The result would be the retrieval of a set of 3D images that are best matched with the input image.