Proceedings Article10.1117/12.2053938
Partially occluded object reconstruction using multiple Kinect sensors
8
TL;DR: This paper proposes a new method for the occluded object visualization using multiple Kinect sensors that can be visualized with a high quality and will not take a long time as other methods.
read more
Abstract: In this paper, we propose a new method for the occluded object visualization using multiple Kinect sensors. The quality of occluded object reconstructed from conventional reconstruction method with elemental images captured by common camera arrays in integral imaging is usually degraded due to the existence of occlusion object in the 3D space which is a common case in reality. Even though some occlusion removal algorithms were proposed to improve the resolution of reconstructed occluded object, all of them are very time-consuming and make the 3D object reconstruction process inefficient. On the contrary, the Kinect sensors not only provide RGB images but also depth images. Since the depth and RGB color image is captured by two different cameras on Kinect sensor at different location, the depth image should be mapped to the color image. After image mapping (or registration), the same pixel location on depth and RGB color image would represent the same 3D space point. As result, the depth image after mapping can be used to remove the occlusion object in the corresponding RGB image (or elemental image) before doing object reconstruction in integral imaging. Consequently, the occluded object can be visualized with a high quality and will not take a long time as other methods.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Fusion of information from multiple Kinect sensors for 3D object reconstruction
Abstract: In this paper, we estimate the accuracy of 3D object reconstruction using multiple Kinect sensors. First, we discuss the calibration of multiple Kinect sensors, and provide an analysis of the accuracy and resolution of the depth data. Next, the precision of coordinate mapping between sensors data for registration of depth and color images is evaluated. We test a proposed system for 3D object reconstruction with four Kinect V2 sensors and present reconstruction accuracy results. Ex-periments and computer simulation are carried out using Matlab and Kinect V2.
29
Three-dimensional integral imaging for gesture recognition under occlusions
Filiberto Pla,Pedro Latorre-Carmona,Eva Salvador-Balaguer,Bahram Javidi +3 more
- 14 May 2018
TL;DR: This paper presents results corresponding to the application of the integral imaging 3D acquisition technique for the recognition of human gestures, when there are occlusions that may hinder the recognition capability, and presents its capability against that given by an RGB-D sensor (Kinect) and that obtained when only one of the cameras in the camera array is used.
Towards 3D Television Through Fusion of Kinect and Integral-Imaging Concepts
Seokmin Hong,Dong-Hak Shin,Byung-Gook Lee,Adrian Dorado,Genaro Saavedra,Manuel Martínez-Corral +5 more
TL;DR: It is demonstrated that with the information provided by a kinect device it is possible to generate an array of microimages ready for their projection onto an integral-imaging monitor.
15
Real-time dense 3D object reconstruction using RGB-D sensor
Alexey Ruchay,Alexey Ruchay,Konstantin Dorofeev,Vsevolod Kalschikov +3 more
- 21 Aug 2020
TL;DR: A new algorithm for dense 3D object reconstruction using a RGB-D sensor at high rate and an efficient merging of the current and incoming point clouds obtained with the Iterative Closest Point is suggested.
5
Duodepth: Static Gesture Recognition Via Dual Depth Sensors
Ilya Chugunov,Avideh Zakhor +1 more
- 21 Sep 2019
TL;DR: In this article, a more classic approach using iterative closest point registration was used to accurately fuse point clouds and a single PointNet architecture for classification, and the other was a dual Point-Net architecture without registration, showing a 39.2% reduction in misclassification for the fused point cloud method and 53.4% for the dual PointNet, when compared to a standard single camera pipeline.
1
References
Advances in three-dimensional integral imaging: sensing, display, and applications [Invited]
TL;DR: This Commemorative Review presents an overview of literature on physical principles and applications of integral imaging, and applications including 3D underwater imaging, 3D imaging in photon-starved environments, 2D tracking of occluded objects,3D optical microscopy, and 3D polarimetric imaging are reviewed.
537
Three-dimensional recognition of occluded objects by using computational integral imaging.
TL;DR: The correlation results of occluded 3D images for volumetric reconstruction show substantial improvements compared with those for conventional 2D imaging of Occluded images.
136
Improved resolution 3D object sensing and recognition using time multiplexed computational integral imaging
Sherif Kishk,Bahram Javidi +1 more
TL;DR: It is shown that the use of a non-stationary microlens array improves the longitudinal distance estimation quantization error and the proposed method overcomes the Nyquist upper limit for the resolution.
128
Real-time three-dimensional object recognition with multiple perspectives imaging.
TL;DR: This 3D processor is the first to apply the principle of integral photography to 3D image recognition and can recognize a slightly out-of-plane rotated 3D object.
90
Fast 3D Computational Integral Imaging Using Graphics Processing Unit
TL;DR: This paper presents a fast three-dimensional (3D) integral imaging system via a graphics processing unit (GPU) which allows parallel processing with multiple processors and shows that it can significantly accelerate 3D scene reconstruction in II using the GPU based stream-processing model.
38