3D image reconstruction: Techniques, applications and challenges

Khilar, Rashmita; Chitrakala, S.; SelvamParvathy, SurenderNath

doi:10.1109/icoiss.2013.6678395

Cited by 18 publications

(7 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In detail, EdgeConv adopts an asymmetric edge function ℎ 𝜃𝜃 (𝑝𝑝 𝑖𝑖 , 𝑝𝑝 𝑗𝑗 ) = ℎ �⃗ 𝜃𝜃 �p 𝑖𝑖 , p 𝑗𝑗 − p 𝑖𝑖 � where the centers of the patch point set p 𝑖𝑖 get the global geometric information and p 𝑗𝑗 − p 𝑖𝑖 capture the local neighborhood point information. Thus, the feature is denoted as shown in Equation (1).…”

Section: A Preliminary: Edge Convolutionmentioning

confidence: 99%

See 1 more Smart Citation

Cascade Convolution Neural Network for Point Set Generation

Lee²

2021

IEEE Access

View full text Add to dashboard Cite

Automatic and efficient 3D object modeling has become critical in industrial applications. The advancement of deep convolutional neural networks (CNNs) has prompted researchers to use CNNs for learning 3D geometry information directly from images. However, the feature maps directly extracted by CNNs are more suitable for image processing tasks because they contain more deep texture information of the entire 2D image. Compared with this, 3D reconstruction tasks using CNNs demand geometric information about a specific object. The existing architecture mainly tries to infer the geometric structure through texture information, which leads to an uneven distribution of points in the generated point cloud object. To address this problem, we propose a cascade point set generation network (CPSGN) that deforms the predicted object while more effectively inferring the object's 3D geometric information from the 2D image so that the distribution of the final object becomes more uniform and denser. The CPSGN consists of a point set generation part for producing a basic 3D object and a point deformation part for fine-tuning the basic 3D object. In addition, we designed a projection loss that optimizes the geometry of the model by measuring shape differences from multiple perspectives. Experimental results on different benchmark datasets indicate that the produced point-based model outperforms existing approaches.

show abstract

Section: A Preliminary: Edge Convolutionmentioning

confidence: 99%

“…The reconstruction of 3D objects from 2D images is attracting much attention for various applications such as mapping in robotics navigation, 3D animation, and 3D printing [1]. According to the information of an input 2D image and the representation of an output 3D object, 3D reconstruction includes multiple approaches.…”

Section: Introductionmentioning

confidence: 99%

Cascade Convolution Neural Network for Point Set Generation

Lee²

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Several solutions have been developed over the years to ensure the reliable reconstruction of a given environment, some adopting traditional image processing [ 1 , 2 , 3 , 4 , 5 ], and some more up-to-date learning approaches [ 6 , 7 ]. In fact, 3D sensing and reconstruction is a necessary building block behind a number of technologies in industry, including robotics, landslide mapping, gaming, mixed reality, archaeology, and medicine, to name a few [ 8 , 9 , 10 ]. Despite the efforts expended by the research community toward providing progressively more accurate models capable of sensing and reconstructing 3D environments, a number of challenges remain.…”

Section: Introductionmentioning

confidence: 99%

Mobile-Based 3D Modeling: An In-Depth Evaluation for the Application in Indoor Scenarios

Pellegrini¹,

Orlandi²,

Sevegnani³

et al. 2021

J. Imaging

View full text Add to dashboard Cite

Indoor environment modeling has become a relevant topic in several application fields, including augmented, virtual, and extended reality. With the digital transformation, many industries have investigated two possibilities: generating detailed models of indoor environments, allowing viewers to navigate through them; and mapping surfaces so as to insert virtual elements into real scenes. The scope of the paper is twofold. We first review the existing state-of-the-art (SoA) of learning-based methods for 3D scene reconstruction based on structure from motion (SFM) that predict depth maps and camera poses from video streams. We then present an extensive evaluation using a recent SoA network, with particular attention on the capability of generalizing on new unseen data of indoor environments. The evaluation was conducted by using the absolute relative (AbsRel) measure of the depth map prediction as the baseline metric.

show abstract

“…Inspired by human eyes, stereo vision uses two cameras displaced horizontally on a calibrated platform to reconstruct three-dimensional (3D) perception of an object based on the visual properties such as shape, color distribution, and illumination [1]- [2]. With the emerging of the imaging sensors and computer vision technology, the application of the stereo vision is extended to various industrial fields including virtual reality, gaming, military training filmmaking, and sports activities [3]- [4].…”

Section: Introductionmentioning

confidence: 99%

Stereo Vision Human Motion Detection and Tracking in Uncontrolled Environment

Seng¹,

Lim²,

Gopal³

et al. 2018

TELKOMNIKA

View full text Add to dashboard Cite

Stereo vision in detecting human motion is an emerging research for automation, rob otics, and sports science field due to the advancement of imaging sensors and information technology. The difficulty of human motion detection and tracking is relatively complex when it is applied to uncontrolled environment. In this paper, a hyb rid filter approach is proposed to detect human motion in the stereo vision. The hyb rid filter approach integrates Gaussian filter and median filter to reduce the coverage of shadow and sudden change of illumination. In addition, sequential thinning and thickening morphological method is used to construct the skeleton model. The proposed hyb rid approach is compared with the normalized filter. As a result, the proposed approach produces b etter skeleton model with less influential effect on shadow and illumination. The output results of the proposed approach can show up to 86% of average accuracy matched with skeleton model. In addition, ob tains approximately 94% of sensitivity measurement in the stereo vision. The proposed approach using hyb rid filter and sequential morphology could improve the performance of the detection in the uncontrolled environment.

show abstract

3D image reconstruction: Techniques, applications and challenges

Cited by 18 publications

References 10 publications

Cascade Convolution Neural Network for Point Set Generation

Cascade Convolution Neural Network for Point Set Generation

Mobile-Based 3D Modeling: An In-Depth Evaluation for the Application in Indoor Scenarios

Stereo Vision Human Motion Detection and Tracking in Uncontrolled Environment

Contact Info

Product

Resources

About