As a 3D sensor and its application technology had been developed and popularized, technical demands for extraction of significant components from spatial models are increasing. Acquiring spatial information from a point cloud is very important and fundamental technique to process a large number of points. To acquire spatial information from the point cloud, we focused on a slant of a normal vector to obtain a geometric shape from the point cloud. In this paper, we propose 3D point cloud cluster analysis based on principal component analysis of normal vectors distribution. Our experimental results showed that the proposed method is robust to noises or spatial holes. Cluster analysis of this normal vector distribution indicated that eigenvalues are affected by a local neighborhood.
Holography is an ultimate three-dimensional (3-D) technology which can record and reconstruct objects with 3-D information. One of the methods that make hologram is to calculate propagation of the light from imaginary object on a computer. We take photo from around the object and perform SFM (Structure From Motion) that refers to the process of estimating 3-D structures from 2-D images. And then, we use the above-mentioned method to make hologram. This paper shows making hologram method, procedure, and actually created hologram.
We previously proposed a wide depth-range (WDR) super multi-view (SMV) display. It can enlarge the display depth range of three-dimensional (3D) images without visual fatigue by combining the multiple SMV displays in the depth direction. The conventional display system could not correspond to the vertical movement of the observer because it provided only horizontal parallax. In this study, the display system was modified to provide pseudo vertical parallax by introducing a face tracking system, and consequently the system can display full-parallax 3D images.
An integral-imaging-based capture system with enhanced horizontal viewing zone by use of a camera array has been developed. We previously proposed a method of enlarging horizontal viewing zone by arranging a rectangle elemental image and tilting a lens array to avoid overlapping the elemental images. In order to capture the elemental images with enlarged horizontal viewing zone, three-dimensional images are captured by use of a lens array and camera array, and the both arrays are tilted in the same way as the lens array of the display system with enhanced horizontal viewing zone by use of the previously proposed method. We conducted an experiment to capture and display three-dimensional images and confirmed the validity of the proposed system.
It is common to display first person view on HMD, which is used for virtual reality and augmented reality. On the other hand, video games use many views not only first person view but also several third person views and these views are already quite common for changing between different views in video games. We need to clarify user operations in various views for future free-viewpoint systems using HMDs. In this study, we performed experiments in three different views and measured the operating times of two tasks using video see-through head mounted display (HMD) for each.