scholarly journals Image-Based Rendering of Cars using Semantic Labels and Approximate Reflection Flow

2020 ◽  
Vol 3 (1) ◽  
pp. 1-17
Author(s):  
Simon Rodriguez ◽  
Siddhant Prakash ◽  
Peter Hedman ◽  
George Drettakis
Keyword(s):  
Initial Approximation ◽  
Visual Quality ◽  
3D Navigation ◽  
New Approach ◽  
Flow Computation ◽  
Image Based Rendering ◽  
Stereo Reconstruction ◽  
Interactive 3D ◽  
Transparent Windows ◽  
Synthesis Solution

Image-Based Rendering (IBR) has made impressive progress towards highly realistic, interactive 3D navigation for many scenes, including cityscapes. However, cars are ubiquitous in such scenes; multi-view stereo reconstruction provides proxy geometry for IBR, but has difficulty with shiny car bodies, and leaves holes in place of reflective, semi-transparent windows on cars. We present a new approach allowing free-viewpoint IBR of cars based on an approximate analytic reflection flow computation on curved windows. Our method has three components: a refinement step of reconstructed car geometry guided by semantic labels, that provides an initial approximation for missing window surfaces and a smooth completed car hull; an efficient reflection flow computation using an ellipsoid approximation of the curved car windows that runs in real-time in a shader and a reflection/background layer synthesis solution. These components allow plausible rendering of reflective, semi-transparent windows in free viewpoint navigation. We show results on several scenes casually captured with a single consumer-level camera, demonstrating plausible car renderings with significant improvement in visual quality over previous methods.

A Framework for Interactive 3D Rendering on Mobile Devices

2014 ◽  
Vol 4 (2) ◽  
pp. 18-31 ◽  
Author(s):  
Ekta Walia ◽  
Vishal Verma
Keyword(s):  
Mobile Devices ◽  
Visual Quality ◽  
Experimental Results ◽  
3D Graphics ◽  
Image Based Rendering ◽  
3D Rendering ◽  
Computing Platform ◽  
Java Platform ◽  
Interactive 3D ◽  
Better Than

Although mobile devices have now become an important computing platform, however most of them still lack hardware accelerated graphics. Therefore, interactive 3D rendering on these devices is a difficult task. This paper describes different approaches for 3D rendering on mobile devices and the associated challenges. It then investigates different solutions to resolve these problems and proposes a framework that uses Image Based Rendering (IBR) technique to render interactive 3D graphics on mobile devices. Further, the performance of proposed framework is compared with Geometry Based Rendering (GBR). The experimental results show that the proposed framework performs better than the geometry-based techniques in terms of rendering time, visual quality and memory requirements. The results also show that the rendering time of the proposed framework is independent of the scene complexity. The experiments are performed in Java Platform Micro Edition (JavaME) environment with Sun JavaME Phone emulator.

Automatic polygon mesh repair and simplification for three-dimensional human modeling

2017 ◽  
Vol 31 (19-21) ◽  
pp. 1740077 ◽  
Author(s):  
Yaolin Zhu ◽  
Li Tian ◽  
Taoruan Wan
Keyword(s):  
Mesh Repair ◽  
Three Dimensional ◽  
Quality Criteria ◽  
Visual Quality ◽  
Triangle Mesh ◽  
New Approach ◽  
Closed Triangle ◽  
Polygon Mesh ◽  
Human Modeling ◽  
Qualitative Measurements

We present a new approach to repair degenerated and self-intersected mesh of digitized human models, which automatically transforms a raw digitized mesh into a single manifold and closed triangle mesh. The algorithm interactively removes growing neighborhoods of undesired elements and fills resulted surface gaps to yield meshes which subject to user-specified quality criteria. We demonstrate our experimental results by multiple sets of low-quality original models with qualitative measurements compared with a number of existing methods. The results show that the proposed method produces high visual quality meshes with minimal changes of the original mesh appearances.

The virtual museum: Interactive 3D navigation of a multimedia database

1992 ◽  
Vol 3 (3) ◽  
pp. 183-197 ◽  
Author(s):  
Gavin Miller ◽  
Eric Hoffert ◽  
Shenchang Eric Chen ◽  
Elizabeth Patterson ◽  
Dean Blackketter ◽  
...  
Keyword(s):  
Multimedia Database ◽  
Virtual Museum ◽  
3D Navigation ◽  
Interactive 3D

scholarly journals Reconstructing conductivities with boundary corrected D-bar method

2014 ◽  
Vol 22 (6) ◽  
Author(s):  
Samuli Siltanen ◽  
Janne P. Tamminen
Keyword(s):  
Numerical Experiments ◽  
Electrical Impedance ◽  
Measurement Data ◽  
Visual Quality ◽  
Previous Approach ◽  
Impedance Tomography ◽  
New Approach ◽  
Conductivity Distribution ◽  
Boundary Measurements ◽  
Ill Posed

AbstractThe aim of electrical impedance tomography is to form an image of the conductivity distribution inside an unknown body using electric boundary measurements. The computation of the image from measurement data is a non-linear ill-posed inverse problem and calls for a special regularized algorithm. One such algorithm, the so-called D-bar method, is improved in this work by introducing new computational steps that remove the so far necessary requirement that the conductivity should be constant near the boundary. The numerical experiments presented suggest two conclusions. First, for most conductivities arising in medical imaging, it seems the previous approach of using a best possible constant near the boundary is sufficient. Second, for conductivities that have high contrast features at the boundary, the new approach produces reconstructions with smaller quantitative error and with better visual quality.

scholarly journals Convergent series for polynomial lattice models with complex actions

2019 ◽  
Vol 34 (30) ◽  
pp. 1950243
Author(s):  
Vasily Sazonov
Keyword(s):  
Monte Carlo ◽  
Lattice Models ◽  
Initial Approximation ◽  
Convergent Series ◽  
Two Dimensional ◽  
Complex Action ◽  
New Approach ◽  
Monte Carlo Techniques ◽  
Sign Problem ◽  
Non Gaussian

Lattice models with complex actions are important for the understanding of matter at finite densities, but not accessible by the standard Monte Carlo techniques due to the sign problem. Here, we propose a new approach aiming to avoid the complex action/sign problem, by extending the method of convergent series (CS) with a non-Gaussian initial approximation. The main properties of the new series are demonstrated on the example of the two-dimensional oscillating integral.

scholarly journals Interactive 3D Navigation System for Image-guided Surgery

2009 ◽  
Vol 8 (1) ◽  
pp. 9-16 ◽  
Author(s):  
Huy Hoang Tran ◽  
Kiyoshi Matsumiya ◽  
Ken Masamune ◽  
Ichiro Sakuma ◽  
Takeyoshi Dohi ◽  
...  
Keyword(s):  
Navigation System ◽  
High Speed ◽  
3D Structure ◽  
Three Dimensional ◽  
Surgical Instruments ◽  
Surface Model ◽  
Surface Rendering ◽  
3D Navigation ◽  
Registration Method ◽  
Interactive 3D

This paper presents a novel surgery navigation system based on a three-dimensional (3D) imaging technique, integral videography (IV). In our system, the 3D structure of the object of interest is reconstructed using surface rendering and corresponding pixel distribution methods. We developed a high-speed algorithm that renders high-quality IV images from the surface model in real time and allows interactions like rotating and scaling to be done smoothly. Using the patient-image registration method, IV images can be displayed with the correct size and relative position with respect to the surgical instruments. Experiments were carried out with various anatomical models, and the results show that our system could be useful in many clinical situations such as orthopedic surgery and neurosurgery.

Predicting Single Observer’s Votes from Objective Measures using Neural Networks

2020 ◽  
Vol 2020 (11) ◽  
pp. 130-1-130-8
Author(s):  
Lohic Fotio Tiotsop ◽  
Tomas Mizdos ◽  
Miroslav Uhrina ◽  
Peter Pocta ◽  
Marcus Barkowsky ◽  
...  
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Numerical Experiments ◽  
Visual Quality ◽  
Objective Measures ◽  
New Approach ◽  
Opinion Score ◽  
The Mean ◽  
Insight Into

The last decades witnessed an increasing number of works aiming at proposing objective measures for media quality assessment, i.e. determining an estimation of the mean opinion score (MOS) of human observers. In this contribution, we investigate a possibility of modeling and predicting single observer’s opinion scores rather than the MOS. More precisely, we attempt to approximate the choice of one single observer by designing a neural network (NN) that is expected to mimic that observer behavior in terms of visual quality perception. Once such NNs (one for each observer) are trained they can be looked at as “virtual observers” as they take as an input information about a sequence and they output the score that the related observer would have given after watching that sequence. This new approach allows to automatically get different opinions regarding the perceived visual quality of a sequence whose quality is under investigation and thus estimate not only the MOS but also a number of other statistical indexes such as, for instance, the standard deviation of the opinions. Large numerical experiments are performed to provide further insight into a suitability of the approach.

scholarly journals A secure image steganography based on JND model

2020 ◽  
Vol 10 (2) ◽  
pp. 2088
Author(s):  
Tuan Duc Nguyen ◽  
Dung Huu Le
Keyword(s):  
Selection Rule ◽  
Visual Quality ◽  
Channel Selection ◽  
Just Noticeable Difference ◽  
Perceptual Quality ◽  
Stego Image ◽  
New Approach ◽  
Matrix Embedding ◽  
Selection Of

Minimizing distortion produced by embedding process is very important to improve the security of hidden message and maintain the high visual quality of stego images. To achieve these objectives, an effective strategy is to perform pixel selection which is well-known as a channel selection rule. In this approach, a pixel associated with the smallest image degradation is chosen to carry secret bits. From these facts, in this paper, a new secure channel selection rule for digital images in spatial domain is designed and proposed. In this new approach, the modified matrix embedding method is utilized as data hiding method because it introduces more than one embedding change to be performed. This enables us to select a suitable pixel to embed message bits with less degradation yielded in a stego-image. In pixel selection of the proposed method, a just noticeable difference value and gradient value of a considering pixel are employed together. The experimental results (which were conducted on 10,000 uncompressed images) indicate that stego images of the proposed approach achieve a higher perceptual quality and security than those of the stego-images created by the previous approaches.

Sign in / Sign up

Export Citation Format

Share Document