Extreme Feature Regions Detection and Accurate Quality Assessment for Point-Cloud 3D Reconstruction

Abstract

The 3-D reconstruction methods based on structure from motion (SfM) pipeline mainly use the traditional scale invariant feature detecting methods for feature matching. This type of methods suffers from the accuracy with affine matching in the image-based modeling system. In this paper, we propose an affine-invariant feature detection and matching method which is accurate and fast based on the three types of critical points in Morse theory called precise extreme feature region (PEFR). We also exploit the evaluation method for the reconstruction results without the ground truth. We propose the average descending rate to quantitatively evaluate the accuracy of the reconstruction model. Using PEFR in an SfM pipeline to build 3-D point-cloud, our method presents better results and less runtime compared to the state-of-the-art approaches. In addition, our quality assessment method presents a practical comparison of the reconstruction result based on the geometric accuracy of the reconstruction models. The experimental results prove that our PEFR method is faster and more accurate and works well on the reconstruction pipeline. The assessment method is fit for the quality assessment on 3-D reconstruction. INDEX TERMS Affine invariant, image matching, accurate quality assessment, extreme feature regions, 3D reconstruction.

Semantic Point Cloud Segmentation Using Fast Deep Neural Network and DCRF

“…Semantic segmentation can then be extended to graph convolution to handle large-scale point clouds [ 17 ] and KD-tree to handle non-uniform point distribution [ 18 , 19 ]. Meanwhile, the registration of point clouds and the detection algorithm of extreme feature areas are also constantly reducing the impact of point cloud noise [ 20 ].…”

A Light Field FDL-HCGH Feature in Scale-Disparity Space

“…Features extracted from higher dimensional data have shown a great potential for improving both matching accuracy and robustness. For example, the authors in [11], [16] and [17] introduce methods for extracting features from 3D point cloud data in order to have a more precise 3D object modeling and recognition. Gao et al [13] propose a planar point feature detection for improving the reconstruction precision of a RGB-D SLAM system, in presence of noise.…”

Energy minimisation‐based multi‐class multi‐instance geometric primitives extraction from 3D point clouds

“…The rise of artificial intelligence promotes the rapid development of three‐dimensional (3D) vision‐based intelligent applications, such as unmanned driving, 3D television‐based immersive interactive media etc. Geometric primitives contained in 3D point clouds can provide not only the meaningful and concise abstraction of 3D data [1] but also the possibility of high‐level interaction for users, which can further improve these 3D vision‐based intelligent applications. However, how to efficiently and robustly extract multiple geometric primitives from point clouds is still a challenge.…”