arXiv subjects

Tobias Ritschel

Publications and source records attributed to Tobias Ritschel.

At least 73 records · Page 4Linked to original sources

arXiv2020

Neural View-Interpolation for Sparse Light Field Video

We suggest representing light field (LF) videos as "one-off" neural networks (NN), i.e., a learned mapping from view-plus-time coordinates to high-resolution color values, trained on sparse views. Initially, this sounds like a bad idea for three main reasons: First, a NN LF will likely have less quality than a same-sized pixel basis representation. Second, only few training data, e.g., 9 exemplars per frame are available for sparse LF videos. Third, there is no generalization across LFs, but across view and time instead. Consequently, a network needs to be trained for each LF video. Surprisingly, these problems can turn into substantial advantages: Other than the linear pixel basis, a NN has to come up with a compact, non-linear i.e., more intelligent, explanation of color, conditioned on the sparse view and time coordinates. As observed for many NN however, this representation now is interpolatable: if the image output for sparse view coordinates is plausible, it is for all intermediate, continuous coordinates as well. Our specific network architecture involves a differentiable occlusion-aware warping step, which leads to a compact set of trainable parameters and consequently fast learning and fast execution.

cs.GR↗

arXiv2020

Learning a Neural 3D Texture Space from 2D Exemplars

We propose a generative model of 2D and 3D natural textures with diversity, visual fidelity and at high computational efficiency. This is enabled by a family of methods that extend ideas from classic stochastic procedural texturing (Perlin noise) to learned, deep, non-linearities. The key idea is a hard-coded, tunable and differentiable step that feeds multiple transformed random 2D or 3D fields into an MLP that can be sampled over infinite domains. Our model encodes all exemplars from a diverse set of textures without a need to be re-trained for each exemplar. Applications include texture interpolation, and learning 3D textures from 2D exemplars.

cs.CV↗

arXiv2020

Collapse of layer dimerization in the photo-induced hidden state of 1T-TaS2

Photo-induced switching between collective quantum states of matter is a fascinating rising field with exciting opportunities for novel technologies. Presently very intensively studied examples in this regard are nanometer-thick single crystals of the layered material 1T-TaS2 , where picosecond laser pulses can trigger a fully reversible insulator-to-metal transition (IMT). This IMT is believed to be connected to the switching between metastable collective quantum states, but the microscopic nature of this so-called hidden quantum state remained largely elusive up to now. Here we determine the latter by means of state-of-the-art x-ray diffraction and show that the laser-driven IMT involves a marked rearrangement of the charge and orbital order in the direction perpendicular to the TaS2-layers. More specifically, we identify the collapse of inter-layer molecular orbital dimers, which are a characteristic feature of the insulating phase, as a key mechanism for the non-thermal IMT in 1T-TaS2, which indeed involves a collective transition between two truly long-range ordered electronic crystals.

cond-mat.str-el↗

arXiv2019

Deep Appearance Maps

We propose a deep representation of appearance, i. e., the relation of color, surface orientation, viewer position, material and illumination. Previous approaches have useddeep learning to extract classic appearance representationsrelating to reflectance model parameters (e. g., Phong) orillumination (e. g., HDR environment maps). We suggest todirectly represent appearance itself as a network we call aDeep Appearance Map (DAM). This is a 4D generalizationover 2D reflectance maps, which held the view direction fixed. First, we show how a DAM can be learned from images or video frames and later be used to synthesize appearance, given new surface orientations and viewer positions. Second, we demonstrate how another network can be used to map from an image or video frames to a DAM network to reproduce this appearance, without using a lengthy optimization such as stochastic gradient descent (learning-to-learn). Finally, we show the example of an appearance estimation-and-segmentation task, mapping from an image showingmultiple materials to multiple deep appearance maps.

cs.CV↗

arXiv2019

Total Denoising: Unsupervised Learning of 3D Point Cloud Cleaning

We show that denoising of 3D point clouds can be learned unsupervised, directly from noisy 3D point cloud data only. This is achieved by extending recent ideas from learning of unsupervised image denoisers to unstructured 3D point clouds. Unsupervised image denoisers operate under the assumption that a noisy pixel observation is a random realization of a distribution around a clean pixel value, which allows appropriate learning on this distribution to eventually converge to the correct value. Regrettably, this assumption is not valid for unstructured points: 3D point clouds are subject to total noise, i. e., deviations in all coordinates, with no reliable pixel grid. Thus, an observation can be the realization of an entire manifold of clean 3D points, which makes a naïve extension of unsupervised image denoisers to 3D point clouds impractical. Overcoming this, we introduce a spatial prior term, that steers converges to the unique closest out of the many possible modes on a manifold. Our results demonstrate unsupervised denoising performance similar to that of supervised learning with clean data when given enough training examples - whereby we do not need any pairs of noisy and clean training data.

cs.CV↗

arXiv2019

Learning to Predict Image-based Rendering Artifacts with Respect to a Hidden Reference Image

Image metrics predict the perceived per-pixel difference between a reference image and its degraded (e. g., re-rendered) version. In several important applications, the reference image is not available and image metrics cannot be applied. We devise a neural network architecture and training procedure that allows predicting the MSE, SSIM or VGG16 image difference from the distorted image alone while the reference is not observed. This is enabled by two insights: The first is to inject sufficiently many un-distorted natural image patches, which can be found in arbitrary amounts and are known to have no perceivable difference to themselves. This avoids false positives. The second is to balance the learning, where it is carefully made sure that all image errors are equally likely, avoiding false negatives. Surprisingly, we observe, that the resulting no-reference metric, subjectively, can even perform better than the reference-based one, as it had to become robust against mis-alignments. We evaluate the effectiveness of our approach in an image-based rendering context, both quantitatively and qualitatively. Finally, we demonstrate two applications which reduce light field capture time and provide guidance for interactive depth adjustment.

cs.GR↗

arXiv2019

Deep-learning the Latent Space of Light Transport

We suggest a method to directly deep-learn light transport, i. e., the mapping from a 3D geometry-illumination-material configuration to a shaded 2D image. While many previous learning methods have employed 2D convolutional neural networks applied to images, we show for the first time that light transport can be learned directly in 3D. The benefit of 3D over 2D is, that the former can also correctly capture illumination effects related to occluded and/or semi-transparent geometry. To learn 3D light transport, we represent the 3D scene as an unstructured 3D point cloud, which is later, during rendering, projected to the 2D output image. Thus, we suggest a two-stage operator comprising of a 3D network that first transforms the point cloud into a latent representation, which is later on projected to the 2D output image using a dedicated 3D-2D network in a second step. We will show that our approach results in improved quality in terms of temporal coherence while retaining most of the computational efficiency of common 2D methods. As a consequence, the proposed two stage-operator serves as a valuable extension to modern deferred shading approaches.

cs.GR↗

arXiv2018

Single-image Tomography: 3D Volumes from 2D Cranial X-Rays

As many different 3D volumes could produce the same 2D x-ray image, inverting this process is challenging. We show that recent deep learning-based convolutional neural networks can solve this task. As the main challenge in learning is the sheer amount of data created when extending the 2D image into a 3D volume, we suggest firstly to learn a coarse, fixed-resolution volume which is then fused in a second step with the input x-ray into a high-resolution volume. To train and validate our approach we introduce a new dataset that comprises of close to half a million computer-simulated 2D x-ray images of 3D volumes scanned from 175 mammalian species. Applications of our approach include stereoscopic rendering of legacy x-ray images, re-rendering of x-rays including changes of illumination, view pose or geometry. Our evaluation includes comparison to previous tomography work, previous learning methods using our data, a user study and application to a set of real x-rays.

cs.GR↗

arXiv2018

Monte Carlo Convolution for Learning on Non-Uniformly Sampled Point Clouds

Deep learning systems extensively use convolution operations to process input data. Though convolution is clearly defined for structured data such as 2D images or 3D volumes, this is not true for other data types such as sparse point clouds. Previous techniques have developed approximations to convolutions for restricted conditions. Unfortunately, their applicability is limited and cannot be used for general point clouds. We propose an efficient and effective method to learn convolutions for non-uniformly sampled point clouds, as they are obtained with modern acquisition techniques. Learning is enabled by four key novelties: first, representing the convolution kernel itself as a multilayer perceptron; second, phrasing convolution as a Monte Carlo integration problem, third, using this notion to combine information from multiple samplings at different levels; and fourth using Poisson disk sampling as a scalable means of hierarchical point cloud learning. The key idea across all these contributions is to guarantee adequate consideration of the underlying non-uniform sample distribution function from a Monte Carlo perspective. To make the proposed concepts applicable to real-world tasks, we furthermore propose an efficient implementation which significantly reduces the GPU memory required during the training process. By employing our method in hierarchical network architectures we can outperform most of the state-of-the-art networks on established point cloud segmentation, classification and normal estimation benchmarks. Furthermore, in contrast to most existing approaches, we also demonstrate the robustness of our method with respect to sampling variations, even when training with uniformly sampled data only. To support the direct application of these concepts, we provide a ready-to-use TensorFlow implementation of these layers at https://github.com/viscom-ulm/MCCNN

cs.CV↗

arXiv2018

End-to-end Sampling Patterns

Sample patterns have many uses in Computer Graphics, ranging from procedural object placement over Monte Carlo image synthesis to non-photorealistic depiction. Their properties such as discrepancy, spectra, anisotropy, or progressiveness have been analyzed extensively. However, designing methods to produce sampling patterns with certain properties can require substantial hand-crafting effort, both in coding, mathematical derivation and compute time. In particular, there is no systematic way to derive the best sampling algorithm for a specific end-task. Tackling this issue, we suggest another level of abstraction: a toolkit to end-to-end optimize over all sampling methods to find the one producing user-prescribed properties such as discrepancy or a spectrum that best fit the end-task. A user simply implements the forward losses and the sampling method is found automatically -- without coding or mathematical derivation -- by making use of back-propagation abilities of modern deep learning frameworks. While this optimization takes long, at deployment time the sampling method is quick to execute as iterated unstructured non-linear filtering using radial basis functions (RBFs) to represent high-dimensional kernels. Several important previous methods are special cases of this approach, which we compare to previous work and demonstrate its usefulness in several typical Computer Graphics applications. Finally, we propose sampling patterns with properties not shown before, such as high-dimensional blue noise with projective properties.

cs.GR↗

arXiv2018

Perceptual Rasterization for Head-mounted Display Image Synthesis

We suggest a rasterization pipeline tailored towards the need of head-mounted displays (HMD), where latency and field-of-view requirements pose new challenges beyond those of traditional desktop displays. Instead of rendering and warping for low latency, or using multiple passes for foveation, we show how both can be produced directly in a single perceptual rasterization pass. We do this with per-fragment ray-casting. This is enabled by derivations of tight space-time-fovea pixel bounds, introducing just enough flexibility for requisite geometric tests, but retaining most of the the simplicity and efficiency of the traditional rasterizaton pipeline. To produce foveated images, we rasterize to an image with spatially varying pixel density. To reduce latency, we extend the image formation model to directly produce "rolling" images where the time at each pixel depends on its display location. Our approach overcomes limitations of warping with respect to disocclusions, object motion and view-dependent shading, as well as geometric aliasing artifacts in other foveated rendering techniques. A set of perceptual user studies demonstrates the efficacy of our approach.

cs.GR↗

arXiv2018

Disrupted orbital order and the pseudo-gap in layered 1T-TaS$_2$

We present a state-of-the-art density functional theory (DFT) study which models crucial features of the partially disordered orbital order stacking in the prototypical layered transition metal dichalcogenide 1T-TaS2 . Our results not only show that DFT models with realistic assumptions about the orbital order perpendicular to the layers yield band structures which agree remarkably well with experiments. They also demonstrate that DFT correctly predicts the formation of an excitation pseudo-gap which is commonly attributed to Mott-Hubbard type electron-electron correlations. These results highlight the importance of interlayer interactions in layered transition metal dichalcogenides and serve as an intriguing example of how disorder within an electronic crystal can give rise to pseudo-gap features.

cond-mat.str-el↗

arXiv2018

Learning on the Edge: Explicit Boundary Handling in CNNs

Convolutional neural networks (CNNs) handle the case where filters extend beyond the image boundary using several heuristics, such as zero, repeat or mean padding. These schemes are applied in an ad-hoc fashion and, being weakly related to the image content and oblivious of the target task, result in low output quality at the boundary. In this paper, we propose a simple and effective improvement that learns the boundary handling itself. At training-time, the network is provided with a separate set of explicit boundary filters. At testing-time, we use these filters which have learned to extrapolate features at the boundary in an optimal way for the specific task. Our extensive evaluation, over a wide range of architectural changes (variations of layers, feature channels, or both), shows how the explicit filters result in improved boundary handling. Consequently, we demonstrate an improvement of 5% to 20% across the board of typical CNN applications (colorization, de-Bayering, optical flow, and disparity estimation).

cs.CV↗

arXiv2017

Joint Material and Illumination Estimation from Photo Sets in the Wild

Faithful manipulation of shape, material, and illumination in 2D Internet images would greatly benefit from a reliable factorization of appearance into material (i.e., diffuse and specular) and illumination (i.e., environment maps). On the one hand, current methods that produce very high fidelity results, typically require controlled settings, expensive devices, or significant manual effort. To the other hand, methods that are automatic and work on 'in the wild' Internet images, often extract only low-frequency lighting or diffuse materials. In this work, we propose to make use of a set of photographs in order to jointly estimate the non-diffuse materials and sharp lighting in an uncontrolled setting. Our key observation is that seeing multiple instances of the same material under different illumination (i.e., environment), and different materials under the same illumination provide valuable constraints that can be exploited to yield a high-quality solution (i.e., specular materials and environment illumination) for all the observed materials and environments. Similar constraints also arise when observing multiple materials in a single environment, or a single material across multiple environments. The core of this approach is an optimization procedure that uses two neural networks that are trained on synthetic images to predict good gradients in parametric space given observation of reflected light. We evaluate our method on a range of synthetic and real examples to generate high-quality estimates, qualitatively compare our results against state-of-the-art alternatives via a user study, and demonstrate photo-consistent image manipulation that is otherwise very challenging to achieve.

cs.GR↗

arXiv2017

What Is Around The Camera?

How much does a single image reveal about the environment it was taken in? In this paper, we investigate how much of that information can be retrieved from a foreground object, combined with the background (i.e. the visible part of the environment). Assuming it is not perfectly diffuse, the foreground object acts as a complexly shaped and far-from-perfect mirror. An additional challenge is that its appearance confounds the light coming from the environment with the unknown materials it is made of. We propose a learning-based approach to predict the environment from multiple reflectance maps that are computed from approximate surface normals. The proposed method allows us to jointly model the statistics of environments and material properties. We train our system from synthesized training data, but demonstrate its applicability to real-world data. Interestingly, our analysis shows that the information obtained from objects made out of multiple materials often is complementary and leads to better performance.

cs.CV↗

arXiv2017

The effect of different in-chain impurities on the magnetic properties of the spin chain compound SrCuO$_2$ probed by NMR

The S=1/2 Heisenberg spin chain compound SrCuO2 doped with different amounts of nickel (Ni), palladium (Pd), zinc (Zn) and cobalt (Co) has been studied by means of Cu nuclear magnetic resonance (NMR). Replacing only a few of the S=1/2 Cu ions with Ni, Pd, Zn or Co has a major impact on the magnetic properties of the spin chain system. In the case of Ni, Pd and Zn an unusual line broadening in the low temperature NMR spectra reveals the existence of an impurity-induced local alternating magnetization (LAM), while exponentially decaying spin-lattice relaxation rates $T_1^{-1}$ towards low temperatures indicate the opening of spin gaps. A distribution of gap magnitudes is proven by a stretched spin-lattice relaxation and a variation of $T_1^{-1}$ within the broad resonance lines. These observations depend strongly on the impurity concentration and therefore can be understood using the model of finite segments of the spin 1/2 antiferromagnetic Heisenberg chain, i.e. pure chain segmentation due to S = 0 impurities. This is surprising for Ni as it was previously assumed to be a magnetic impurity with S = 1 which is screened by the neighboring copper spins. In order to confirm the S = 0 state of the Ni, we performed x-ray absorption spectroscopy (XAS) and compared the measurements to simulated XAS spectra based on multiplet ligand-field theory. Furthermore, Zn doping leads to much smaller effects on both the NMR spectra and the spin-lattice relaxation rates, indicating that Zn avoids occupying Cu sites. For magnetic Co impurities, $T_1^{-1}$ does not obey the gap like decrease, and the low-temperature spectra get very broad. This could be related to the increase of the Neel temperature which was observed by recent muSR and susceptibility measurements, and is most likely an effect of the impurity spin $S\neq0$.

cond-mat.str-el↗

arXiv2017

Plausible Shading Decomposition For Layered Photo Retouching

Photographers routinely compose multiple manipulated photos of the same scene (layers) into a single image, which is better than any individual photo could be alone. Similarly, 3D artists set up rendering systems to produce layered images to contain only individual aspects of the light transport, which are composed into the final result in post-production. Regrettably, both approaches either take considerable time to capture, or remain limited to synthetic scenes. In this paper, we suggest a system to allow decomposing a single image into a plausible shading decomposition (PSD) that approximates effects such as shadow, diffuse illumination, albedo, and specular shading. This decomposition can then be manipulated in any off-the-shelf image manipulation software and recomposited back. We perform such a decomposition by learning a convolutional neural network trained using synthetic data. We demonstrate the effectiveness of our decomposition on synthetic (i.e., rendered) and real data (i.e., photographs), and use them for common photo manipulation, which are nearly impossible to perform otherwise from single images.

cs.GR↗

arXiv2016

Novel Views of Objects from a Single Image

Taking an image of an object is at its core a lossy process. The rich information about the three-dimensional structure of the world is flattened to an image plane and decisions such as viewpoint and camera parameters are final and not easily revertible. As a consequence, possibilities of changing viewpoint are limited. Given a single image depicting an object, novel-view synthesis is the task of generating new images that render the object from a different viewpoint than the one given. The main difficulty is to synthesize the parts that are disoccluded; disocclusion occurs when parts of an object are hidden by the object itself under a specific viewpoint. In this work, we show how to improve novel-view synthesis by making use of the correlations observed in 3D models and applying them to new image instances. We propose a technique to use the structural information extracted from a 3D model that matches the image object in terms of viewpoint and shape. For the latter part, we propose an efficient 2D-to-3D alignment method that associates precisely the image appearance with the 3D model geometry with minimal user interaction. Our technique is able to simulate plausible viewpoint changes for a variety of object classes within seconds. Additionally, we show that our synthesized images can be used as additional training data that improves the performance of standard object detectors.

cs.CV↗