scholarly journals Free-viewpoint Indoor Neural Relighting from Multi-view Stereo

2021 ◽  
Vol 40 (5) ◽  
pp. 1-18
Author(s):  
Julien Philip ◽  
Sébastien Morgenthaler ◽  
Michaël Gharbi ◽  
George Drettakis
Keyword(s):  
Three Dimensional ◽  
Mirror Reflection ◽  
Light Sources ◽  
Feature Maps ◽  
Convolutional Network ◽  
Implicit Representation ◽  
Input Feature ◽  
Physically Based ◽  
Indoor Scenes ◽  
Physically Based Rendering

We introduce a neural relighting algorithm for captured indoors scenes, that allows interactive free-viewpoint navigation. Our method allows illumination to be changed synthetically, while coherently rendering cast shadows and complex glossy materials. We start with multiple images of the scene and a three-dimensional mesh obtained by multi-view stereo (MVS) reconstruction. We assume that lighting is well explained as the sum of a view-independent diffuse component and a view-dependent glossy term concentrated around the mirror reflection direction. We design a convolutional network around input feature maps that facilitate learning of an implicit representation of scene materials and illumination, enabling both relighting and free-viewpoint navigation. We generate these input maps by exploiting the best elements of both image-based and physically based rendering. We sample the input views to estimate diffuse scene irradiance, and compute the new illumination caused by user-specified light sources using path tracing. To facilitate the network's understanding of materials and synthesize plausible glossy reflections, we reproject the views and compute mirror images . We train the network on a synthetic dataset where each scene is also reconstructed with MVS. We show results of our algorithm relighting real indoor scenes and performing free-viewpoint navigation with complex and realistic glossy reflections, which so far remained out of reach for view-synthesis techniques.

scholarly journals Coronary artery segmentation in angiographic videos utilizing spatial-temporal information

2020 ◽  
Vol 20 (1) ◽  
Author(s):  
Lu Wang ◽  
Dongxue Liang ◽  
Xiaolei Yin ◽  
Jing Qiu ◽  
Zhiyun Yang ◽  
...  
Keyword(s):  
Coronary Artery ◽  
Blood Vessels ◽  
Three Dimensional ◽  
Input Image ◽  
Medical Image Segmentation ◽  
Temporal Information ◽  
Feature Maps ◽  
Convolutional Network ◽  
Input Layer ◽  
Coronary Artery Segmentation

Abstract Background Coronary artery angiography is an indispensable assistive technique for cardiac interventional surgery. Segmentation and extraction of blood vessels from coronary angiographic images or videos are very essential prerequisites for physicians to locate, assess and diagnose the plaques and stenosis in blood vessels. Methods This article proposes a novel coronary artery segmentation framework that combines a three–dimensional (3D) convolutional input layer and a two–dimensional (2D) convolutional network. Instead of a single input image in the previous medical image segmentation applications, our framework accepts a sequence of coronary angiographic images as input, and outputs the clearest mask of segmentation result. The 3D input layer leverages the temporal information in the image sequence, and fuses the multiple images into more comprehensive 2D feature maps. The 2D convolutional network implements down–sampling encoders, up–sampling decoders, bottle–neck modules, and skip connections to accomplish the segmentation task. Results The spatial–temporal model of this article obtains good segmentation results despite the poor quality of coronary angiographic video sequences, and outperforms the state–of–the–art techniques. Conclusions The results justify that making full use of the spatial and temporal information in the image sequences will promote the analysis and understanding of the images in videos.

scholarly journals Blowing and drifting snow in Alpine terrain: numerical simulation and related field measurements

1998 ◽  
Vol 26 ◽  
pp. 174-178 ◽  
Author(s):  
Peter Gauer
Keyword(s):  
Numerical Simulation ◽  
Numerical Model ◽  
Three Dimensional ◽  
Field Measurements ◽  
Blowing Snow ◽  
Related Field ◽  
Drifting Snow ◽  
Physically Based ◽  
Snow Transport

A physically based numerical model of drifting and blowing snow in three-dimensional terrain is developed. The model includes snow transport by saltation and suspension. As an example, a numerical simulation for an Alpine ridge is presented and compared with field measurements.

Creep Simulation of the Hydroprocessing Reactor Using a Physically-Based CDM Model

2015 ◽  
Author(s):  
Yu Zhou ◽  
Chen Xuedong ◽  
Fan Zhichao ◽  
Jie Dong
Keyword(s):  
Damage Mechanics ◽  
Failure Modes ◽  
Elevated Temperatures ◽  
Three Dimensional ◽  
Critical Issue ◽  
Fe Modelling ◽  
Creep Failure ◽  
Good Tool ◽  
Element Analysis ◽  
Physically Based

Creep failure is one of the most important failure modes in the design of hydroprocessing reactors at elevated temperatures, and the accurate prediction of the creep behavior in structural discontinuities is a critical issue for component design. A physically-based continnum damage mechanics (CDM) model was adopted to describe all three creep stages of 2.25Cr-1Mo-0.25V ferritic steel widely used in manufacturing modern hydroprocessing reactors. The material constants in the damage constitutive equations were identified using an efficient optimization scheme based on genetic algorithm (GA). The user-defined subroutine implementing the CDM model was developed using user programmable features (UPFs) in ANSYS. Three-dimensional finite element analysis of the hydroprocessing reactor was conducted to determine the critical regions, and the studies on the stress redistribution and the prediction of damage evolution in these regions during creep were carried out. The results show that FE modelling based on CDM theory can provide a good tool for creep design of complex engineering components.

Misidentifications

1997 ◽  
Vol 8 (S3) ◽  
pp. 393-397 ◽  
Author(s):  
Alistair Burns
Keyword(s):  
Dimensional Space ◽  
Three Dimensional ◽  
Mirror Reflection ◽  
Three Dimensional Space

Misidentifications are misperceptions (i.e., a form of illusion) with an associated belief or elaboration that is held with delusional intensity. Although misidentifications have been defined in several ways, four main types have been described: (a) presence of persons in the patient's own house (the phantom boarder syndrome); (b) misidentification of the patient's own self (often seen as a misrecognition of his or her own mirror reflection); (c) misidentification of other persons; and (d) misidentification of events on television (the patient imagines these events are occurring in real three-dimensional space).

Development of a three-dimensional color rendition space for tunable solid-state light sources

2021 ◽  
Vol 2021 (29) ◽  
pp. 136-140
Author(s):  
Dorukalp Durmus
Keyword(s):  
Solid State ◽  
Dimensional Space ◽  
Three Dimensional ◽  
Color Discrimination ◽  
Continuous Measure ◽  
Light Sources ◽  
Vast Number ◽  
Lighting Systems ◽  
Single Number

The quality of building electric lighting systems can be assessed using color rendition metrics. However, color rendition metrics are limited in quantifying tunable solid-state light sources, since tunable lighting systems can generate a vast number of different white light spectra, providing flexibility in terms of color quality and energy efficiency. Previous research suggests that color rendition is multi-dimensional in nature, and it cannot be simplified to a single number. Color shifts under a test light source in comparison to a reference illuminant, changes in color gamut, and color discrimination are important dimensions of the quality of electric light sources, which are not captured by a single-numbered metric. To address the challenges in color rendition characterization of modern solid-state light sources, the development of a multi-dimensional color rendition space is proposed. The proposed continuous measure can quantify the change in color rendition ability of tunable solid-state light devices with caveats. Future work, discretization of the continuous color rendition space, will be carried out to address the shortcomings of a continuous three-dimensional space.

Simulation of microstructure evolution during recrystallization using a high resolution three dimensional cellular automaton

2004 ◽  
Vol 120 ◽  
pp. 225-230
Author(s):  
P. Mukhopadhyay ◽  
M. Loeck ◽  
G. Gottstein
Keyword(s):  
High Resolution ◽  
Cellular Automata ◽  
Cellular Automaton ◽  
Microstructure Evolution ◽  
Static Recrystallization ◽  
Three Dimensional ◽  
Computation Time ◽  
Recrystallization Process ◽  
Recovery Condition ◽  
Physically Based

A more refined 3D cellular Automata (CA) algorithm has been developed which has increased the resolution of the space and reduced the computation time and can take care of the complexity of recrystallization process through physically based solutions. This model includes recovery, condition for nucleation and orientation dependent variable nuclei growth as a process of primary static recrystallization. Incorporation of microchemistry effects makes this model suitable for simulating recrystallization behaviour in terms of texture, kinetics and microstructure of different alloys. The model is flexible to couple up with other simulation programs on a common database.

scholarly journals A MultiAgent System for Physically Based Rendering Optimization

2007 ◽  
pp. 149-163 ◽  
Author(s):  
Carlos Gonzalez-Morcillo ◽  
Gerhard Weiss ◽  
Luis Jimenez ◽  
David Vallejo ◽  
Javier Albusac
Keyword(s):  
Multiagent System ◽  
Physically Based ◽  
Physically Based Rendering

Improvement and validation of a physically based model for the shear and transverse crushing of orthotropic composites

2018 ◽  
Vol 53 (12) ◽  
pp. 1681-1696 ◽  
Author(s):  
Sérgio Costa ◽  
Thomas Bru ◽  
Robin Olsson ◽  
André Portugal
Keyword(s):  
Current Model ◽  
Three Dimensional ◽  
Fracture Plane ◽  
Damage Growth ◽  
Pressure Dependency ◽  
Orthotropic Composites ◽  
Fabric Composites ◽  
Final Failure ◽  
Physically Based ◽  
Nonlinear Shear

This paper details a complete crush model for composite materials with focus on shear dominated crushing under a three-dimensional stress state. The damage evolution laws and final failure strain conditions are based on data extracted from shear experiments. The main advantages of the current model include the following: no need to measure the fracture toughness in shear and transverse compression, mesh objectivity without the need for a regular mesh and finite element characteristic length, a pressure dependency of the nonlinear shear response, accounting for load reversal and some orthotropic effects (making the model suitable for noncrimp fabric composites). The model is validated against a range of relevant experiments, namely a through-the-thickness compression specimen and a flat crush coupon with the fibres oriented at 45° and 90° to the load. Damage growth mechanisms, orientation of the fracture plane, nonlinear evolution of Poisson's ratio and energy absorption are accurately predicted.

scholarly journals Deep CNNs with Robust LBP Guiding Pooling for Face Recognition

Sensors ◽  
2018 ◽  
Vol 18 (11) ◽  
pp. 3876 ◽  
Author(s):  
Zhongjian Ma ◽  
Yuanyuan Ding ◽  
Baoqing Li ◽  
Xiaobing Yuan
Keyword(s):  
Neural Networks ◽  
Face Recognition ◽  
Computational Complexity ◽  
Local Binary Pattern ◽  
Weighted Average ◽  
Sliding Window ◽  
Performance Gain ◽  
Feature Maps ◽  
Input Feature ◽  
Noise Impact

Pooling layer in Convolutional Neural Networks (CNNs) is designed to reduce dimensions and computational complexity. Unfortunately, CNN is easily disturbed by noise in images when extracting features from input images. The traditional pooling layer directly samples the input feature maps without considering whether they are affected by noise, which brings about accumulated noise in the subsequent feature maps as well as undesirable network outputs. To address this issue, a robust Local Binary Pattern (LBP) Guiding Pooling (G-RLBP) mechanism is proposed in this paper to down sample the input feature maps and lower the noise impact simultaneously. The proposed G-RLBP method calculates the weighted average of all pixels in the sliding window of this pooling layer as the final results based on their corresponding probabilities of being affected by noise, thus lowers the noise impact from input images at the first several layers of the CNNs. The experimental results show that the carefully designed G-RLBP layer can successfully lower the noise impact and improve the recognition rates of the CNN models over the traditional pooling layer. The performance gain of the G-RLBP is quite remarkable when the images are severely affected by noise.

Sign in / Sign up

Export Citation Format

Share Document