This paper will discuss challenges regarding set design, editing, sound design, rendering, and cinematography for adapting animated short movies to the ScreenX format, Furthermore, we will explore the cinematic elemen...
详细信息
ISBN:
(纸本)9798350374490;9798350374506
This paper will discuss challenges regarding set design, editing, sound design, rendering, and cinematography for adapting animated short movies to the ScreenX format, Furthermore, we will explore the cinematic elements of ScreenX for 3D animations. ScreenX is a novel immersive 270-degree panoramic film format that extends the main cinema screen onto two side walls. Since the introduction of the technology in 2012, only a few animated films have been produced. This expanded cinema system is still in its very early stages. Therefore, more is needed to know about the potential for animated movies. We will introduce a case study and provide guidelines for future productions.
We propose a single-rate method for geometry compression of triangle meshes based on using a neural predictor to predict the encoded vertex positions using connectivity and an already known part of the geometry. The m...
详细信息
ISBN:
(纸本)9783031637483;9783031637490
We propose a single-rate method for geometry compression of triangle meshes based on using a neural predictor to predict the encoded vertex positions using connectivity and an already known part of the geometry. The method is based on standard traversal-based methods but uses a neural predictor for prediction instead of a hand-crafted prediction scheme. The parameters of the neural predictor are learned on a dataset of existing triangle meshes. The method additionally includes an estimate of the prediction uncertainty, which is used to guide the encoding traversal of the mesh. The results of the proposed method are compared with a benchmark method on the ABC dataset using both mechanistic and perceptual metrics.
Traditional 3D content creation tools empower users to bring their imagination to life by giving them direct control over a scene's geometry, appearance, motion, and camera path. Creating computer-generated videos...
详细信息
ISBN:
(纸本)9798350353013;9798350353006
Traditional 3D content creation tools empower users to bring their imagination to life by giving them direct control over a scene's geometry, appearance, motion, and camera path. Creating computer-generated videos, however, is a tedious manual process, which can be automated by emerging text-to-video diffusion models. Despite great promise, video diffusion models are difficult to control, hindering a user to apply their own creativity rather than amplifying it. To address this challenge, we present a novel approach that combines the controllability of dynamic 3D meshes with the expressivity and editability of emerging diffusion models. For this purpose, our approach takes an animated, low-fidelity rendered mesh as input and injects the ground truth correspondence information obtained from the dynamic mesh into various stages of a pre-trained text-to-image generation model to output high-quality and temporally consistent frames. We demonstrate our approach on various examples where motion can be obtained by animating rigged assets or changing the camera path. Project page: ***/generative_rendering.
Diffusion-based methods have achieved prominent success in generating 2D media. However, accomplishing similar proficiencies for scene-level mesh texturing in 3D spatial applications, e.g., XR/VR, remains constrained,...
详细信息
ISBN:
(纸本)9798350374025;9798350374032
Diffusion-based methods have achieved prominent success in generating 2D media. However, accomplishing similar proficiencies for scene-level mesh texturing in 3D spatial applications, e.g., XR/VR, remains constrained, primarily due to the intricate nature of 3D geometry and the necessity for immersive free-viewpoint rendering. In this paper, we propose a novel indoor scene texturing framework, which delivers text-driven texture generation with enchanting details and authentic spatial coherence. The key insight is to first imagine a stylized 360. panoramic texture from the central viewpoint of the scene, and then propagate it to the rest areas with inpainting and imitating techniques. To ensure meaningful and aligned textures to the scene, we develop a novel coarse-to-fine panoramic texture generation approach with dual texture alignment, which both considers the geometry and texture cues of the captured scenes. To survive cluttered geometries during texture propagation, we design a separated strategy, which conducts texture inpainting in visible regions and then learns an implicit imitating network to synthesize textures in occluded and tiny structural areas. Extensive experiments and the immersive VR application on real-world indoor scenes demonstrate the high quality of the generated textures and the engaging experience on VR headsets. Project webpage: https://***/publication/dreamspace.
This study presents the initial design and evaluation of an immersive virtual environment aimed at enhancing the comprehension and learning of 3D transformations in computer graphics. Traditional teaching approaches m...
详细信息
ISBN:
(纸本)9798400711770
This study presents the initial design and evaluation of an immersive virtual environment aimed at enhancing the comprehension and learning of 3D transformations in computer graphics. Traditional teaching approaches may limit spatial understanding by using static 2D representations. To address this, our virtual environment includes interactive modules that let students view and work with three-dimensional objects, allowing them to see how various transformations affect objects in real-time.
CG images are rendered using perspective projections. However, we sometimes experience discomfort while observing these images. The reason is that the perspective projection images do not represent the visual impressi...
详细信息
ISBN:
(数字)9781510679931
ISBN:
(纸本)9781510679931;9781510679924
CG images are rendered using perspective projections. However, we sometimes experience discomfort while observing these images. The reason is that the perspective projection images do not represent the visual impression of real space. Therefore, human visual characteristics can contribute to obtaining more realistic CG. Thus, a magnification rate function was proposed to represent the relationship between the subjective perception of the size and viewing distance of the object to be rendered. The image to which the magnification function was applied was closer to the impression of real space than the perspective projection image. In this study, we propose a method for generating images by applying a magnification rate function to panoramic images. This method enables the reproduction of the impression a person feels in real space as an image, even if it is a panoramic image.
Recently, neural network-based approaches for hologram generation and compression have gained popularity as they allow for efficient inference on GPUs without the need for iterative optimization required in traditiona...
详细信息
ISBN:
(纸本)9798350374025;9798350374032
Recently, neural network-based approaches for hologram generation and compression have gained popularity as they allow for efficient inference on GPUs without the need for iterative optimization required in traditional methods. In this paper, we introduce Neural Holographic Video Compression (NHVC), an end-to-end trainable and scalable model designed for high-quality phase hologram video generation and compression. NHVC consists of an auto-encoder-based phase hologram generator, a latent coder and- two hyper-prior coders. For each input image, the latent features are extracted through the encoder part of the phase generator and then entropy coded at the shared latent coder based on the hyper-prior information. The two hyper-prior coders employ a spatial and a spatio-temporal entropy model for I-frames and P-frames, respectively. With this architecture, our NHVC can offer task-scalability, allowing a single trained model to serve as a phase hologram generator, phase hologram image compressor, or phase hologram video compressor as required. Experimental results on phase hologram video compression with UVG dataset show that our model outperforms 'HoloNet + VVC' by 75.6% BD-Rate reduction, with modest 2K encoding and decoding speeds (5 fps and 12 fps, respectively). For the phase hologram video generation task, our model showed much higher-quality (almost 42dB PSNR) reconstruction using the UVG dataset, while the previous neural generation model HoloNet provides at most 36dB reconstruction quality. We also provide an extensive experimental study on several important design questions such as the need for quadruple extension (QE) in the neural compression model, the feasibility of motion estimation in the phase domain, and an alternative, the need for increasing receptive field to learn better phase features, and variable rate support with a single trained model. It is noteworthy that our model is the first and best neural phase video compression model providing such high
In image morphing applications, deformation is not desired since it looks strange. To avoid deformation, the angles must be locally preserved. By deforming features, we can obtain funny-looking images, however, the sh...
详细信息
ISBN:
(纸本)9798350343557
In image morphing applications, deformation is not desired since it looks strange. To avoid deformation, the angles must be locally preserved. By deforming features, we can obtain funny-looking images, however, the shapes of the features will not look preserved if angles are not preserved. The angle-preserving operations are called conformal maps and locally they expect to be reduced into a concatenation of rotation, uniform scaling, and translation operations. This conceptual framework appears to give us a straightforward framework to construct general conformal maps. We can simply define desired rotation, uniform scaling, and translation in some select points and interpolate them by preserving their conformal property. Unfortunately, this does not work if these local operations are defined by using 3x3 matrices. Our approach is based on the well-known conformal property of maps in complex domains. Namely, any map in a complex domain is angle-preserving. Based on this well-known property of complex numbers, we show that if the local operations are defined as affine operations on complex numbers then any interpolation on complex numbers will give us conformal mapping of the whole domain. This is very useful since we can make extreme exaggeration of some local features and the resulting maps are still guaranteed to be conformal. Using this property, we have implemented an extended and conformal version of feature-based image metamorphosis.
Physics-based simulations are a powerful tool in both computer graphics and engineering applications. Implicit discretization is essential for accurate, stable, and efficient simulations of solids and *** this thesis,...
详细信息
Physics-based simulations are a powerful tool in both computer graphics and engineering applications. Implicit discretization is essential for accurate, stable, and efficient simulations of solids and *** this thesis, we first present a novel implicit Material Point Method (MPM) discretization of spatially varying surface energies. Our discretization is based on surface energy, enabling implicit time stepping and capturing surface gradients without explicitly resolving them as in traction-condition-based approaches. We include an implicit discretization of thermomechanical material coupling with novel particle-based enforcement of Robin boundary conditions. Lastly, we design a particle resampling approach for perfect conservations of linear and angular momentum with Affine-Particle-In-Cell (APIC) [JSS15].The second part presents a novel deep-learning approach to approximate the solution of large, sparse, symmetric, positive-definite linear systems of equations. Our method is motivated by the conjugate gradients algorithm that iteratively selects search directions for minimizing the matrix norm of the approximation error. We use a deep neural network to accelerate convergence via data-driven improvement of the search direction at each iteration. We demonstrate the efficacy of our approach on discretized Poisson equations with millions of degrees of freedom. Our algorithm can reduce the linear system residual to the target tolerance in a small number of iterations, independent of the problem size, and generalize effectively to various systems beyond those encountered during ***, we present improvements to Position Based Dynamics (PBD) [MHH07] and Extended Position Based Dynamics (XPBD) [MMC16] methods, which are variants of implicit time integrator. PBD/XPBD are powerful methods for the real-time simulation of elastic objects, but they do not always converge. We isolate the root cause in the approximate linearization of the nonlinear backward Euler
Locating neck-like features, or locally narrow parts, of a surface is crucial in various applications such as segmentation, shape analysis, path planning, and robotics. Topological methods are often utilized to find t...
详细信息
Locating neck-like features, or locally narrow parts, of a surface is crucial in various applications such as segmentation, shape analysis, path planning, and robotics. Topological methods are often utilized to find the set of shortest loops around handles and tunnels. However, there are abundant neck-like features on genus-0 shapes without any handles. While 3D geometry-aware topological approaches exist to find such loops, their construction can be cumbersome and may even lead to unintuitive loops. Here we present two methods for efficiently computing a complete set of surface loops that are not limited to the topologically nontrivial independent *** first approach is an efficient “topology-aware geometric” method to compute the tightest loops around neck features on surfaces, including genus-0 surfaces. We use the critical points of a processed distance function (such as Morse function) to find both the location and evaluate the significance of possible neck-like features. Critical points of a Morse function defined on a volume provide rich topological and geometric information about the structure of the shape. Our algorithm starts with a volumetric representation of an input surface and then calculates the distance function of mesh points to the boundary surface as a Morse function. We directly create a cutting plane through each neck feature. Each resulting loop can then be tightened to form a closed geodesic representation of the neck feature. Moreover, we offer criteria to measure the significance of a neck feature through the evolution of critical points during the smoothing of the distance function. Furthermore, we speed up the detection process through mesh simplification without compromising the quality of the output *** is known that reducing the dimension of a problem typically boosts efficiency drastically. Hence, we propose our second approach, which is a novel, efficient method that uses the skeleton of the shape to compute surface loops.
暂无评论