View-dependent Appearance Research Articles

NeRF-Texture: Synthesizing Neural Radiance Field Textures.

Texture synthesis is a fundamental problem in computer graphics that would benefit various applications. Existing methods are effective in handling 2D image textures. In contrast, many real-world textures contain meso-structure in the 3D geometry space, such as grass, leaves, and fabrics, which cannot be effectively modeled using only 2D image textures. We propose a novel texture synthesis method with Neural Radiance Fields (NeRF) to capture and synthesize textures from given multi-view images. In the proposed NeRF texture representation, a scene with fine geometric details is disentangled into the meso-structure textures and the underlying base shape. This allows textures with meso-structure to be effectively learned as latent features situated on the base shape, which are fed into a NeRF decoder trained simultaneously to represent the rich view-dependent appearance. Using this implicit representation, we can synthesize NeRF-based textures through patch matching of latent features. However, inconsistencies between the metrics of the reconstructed content space and the latent feature space may compromise the synthesis quality. To enhance matching performance, we further regularize the distribution of latent features by incorporating a clustering constraint. In addition to generating NeRF textures over a planar domain, our method can also synthesize NeRF textures over curved surfaces, which are practically useful. Experimental results and evaluations demonstrate the effectiveness of our approach.

Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields.

Neural Radiance Fields (NeRF) is a popular view synthesis technique that represents a scene as a continuous volumetric function, parameterized by multilayer perceptrons that provide the volume density and view-dependent emitted radiance at each location. While NeRF-based techniques excel at representing fine geometric structures with smoothly varying view-dependent appearance, they often fail to accurately capture and reproduce the appearance of glossy surfaces. We address this limitation by introducing Ref-NeRF, which replaces NeRF's parameterization of view-dependent outgoing radiance with a representation of reflected radiance and structures this function using a collection of spatially-varying scene properties. We show that together with a regularizer on normal vectors, our model significantly improves the realism and accuracy of specular reflections. Furthermore, we show that our model's internal representation of outgoing radiance is interpretable and useful for scene editing.

Baking Neural Radiance Fields for Real-Time View Synthesis.

Neural volumetric representations such as Neural Radiance Fields (NeRF) have emerged as a compelling technique for learning to represent 3D scenes from images with the goal of rendering photorealistic images of the scene from unobserved viewpoints. However, NeRF's computational requirements are prohibitive for real-time applications: rendering views from a trained NeRF requires querying a multilayer perceptron (MLP) hundreds of times per ray. We present a method to train a NeRF, then precompute and store (i.e. "bake") it as a novel representation called a Sparse Neural Radiance Grid (SNeRG) that enables real-time rendering on commodity hardware. To achieve this, we introduce 1) a reformulation of NeRF's architecture, and 2) a sparse voxel grid representation with learned feature vectors. The resulting scene representation retains NeRF's ability to render fine geometric details and view-dependent appearance, is compact (averaging less than 90MB per scene), and can be rendered in real-time (higher than 30 frames per second on a laptop GPU). Actual screen captures are shown in our video.

Neural Radiance Fields From Sparse RGB-D Images for High-Quality View Synthesis.

The recently proposed neural radiance fields (NeRF) use a continuous function formulated as a multi-layer perceptron (MLP) to model the appearance and geometry of a 3D scene. This enables realistic synthesis of novel views, even for scenes with view dependent appearance. Many follow-up works have since extended NeRFs in different ways. However, a fundamental restriction of the method remains that it requires a large number of images captured from densely placed viewpoints for high-quality synthesis and the quality of the results quickly degrades when the number of captured views is insufficient. To address this problem, we propose a novel NeRF-based framework capable of high-quality view synthesis using only a sparse set of RGB-D images, which can be easily captured using cameras and LiDAR sensors on current consumer devices. First, a geometric proxy of the scene is reconstructed from the captured RGB-D images. Renderings of the reconstructed scene along with precise camera parameters can then be used to pre-train a network. Finally, the network is fine-tuned with a small number of real captured images. We further introduce a patch discriminator to supervise the network under novel views during fine-tuning, as well as a 3D color prior to improve synthesis quality. We demonstrate that our method can generate arbitrary novel views of a 3D scene from as few as 6 RGB-D images. Extensive experiments show the improvements of our method compared with the existing NeRF-based methods, including approaches that also aim to reduce the number of input images.

Scalable neural indoor scene rendering

We propose a scalable neural scene reconstruction and rendering method to support distributed training and interactive rendering of large indoor scenes. Our representation is based on tiles. Tile appearances are trained in parallel through a background sampling strategy that augments each tile with distant scene information via a proxy global mesh. Each tile has two low-capacity MLPs: one for view-independent appearance (diffuse color and shading) and one for view-dependent appearance (specular highlights, reflections). We leverage the phenomena that complex view-dependent scene reflections can be attributed to virtual lights underneath surfaces at the total ray distance to the source. This lets us handle sparse samplings of the input scene where reflection highlights do not always appear consistently in input images. We show interactive free-viewpoint rendering results from five scenes, one of which covers an area of more than 100 m 2 . Experimental results show that our method produces higher-quality renderings than a single large-capacity MLP and five recent neural proxy-geometry and voxel-based baseline methods. Our code and data are available at project webpage https://xchaowu.github.io/papers/scalable-nisr.

Neural actor

We propose Neural Actor (NA), a new method for high-quality synthesis of humans from arbitrary viewpoints and under arbitrary controllable poses. Our method is developed upon recent neural scene representation and rendering works which learn representations of geometry and appearance from only 2D images. While existing works demonstrated compelling rendering of static scenes and playback of dynamic scenes, photo-realistic reconstruction and rendering of humans with neural implicit methods, in particular under user-controlled novel poses, is still difficult. To address this problem, we utilize a coarse body model as a proxy to unwarp the surrounding 3D space into a canonical pose. A neural radiance field learns pose-dependent geometric deformations and pose- and view-dependent appearance effects in the canonical space from multi-view video input. To synthesize novel views of high-fidelity dynamic geometry and appearance, NA leverages 2D texture maps defined on the body model as latent variables for predicting residual deformations and the dynamic appearance. Experiments demonstrate that our method achieves better quality than the state-of-the-arts on playback as well as novel pose synthesis, and can even generalize well to new poses that starkly differ from the training poses. Furthermore, our method also supports shape control on the free-view synthesis of human actors.

Refractive Light-Field Features for Curved Transparent Objects in Structure From Motion

Curved refractive objects are common in the human environment, and have a complex visual appearance that can cause robotic vision algorithms to fail. Light-field cameras allow us to address this challenge by capturing the view-dependent appearance of such objects in a single exposure. We propose a novel image feature for light fields that detects and describes the patterns of light refracted through curved transparent objects. We derive characteristic points based on these features allowing them to be used in place of conventional 2D features. Using our features, we demonstrate improved structure-from-motion performance in challenging scenes containing refractive objects, including quantitative evaluations that show improved camera pose estimates and 3D reconstructions. Additionally, our methods converge 15-35% more frequently than the state-of-the-art. Our method is a critical step towards allowing robots to operate around refractive objects, with applications in manufacturing, quality assurance, pick-and-place, and domestic robots working with acrylic, glass and other transparent materials.

Driving-signal aware full-body avatars

We present a learning-based method for building driving-signal aware full-body avatars. Our model is a conditional variational autoencoder that can be animated with incomplete driving signals, such as human pose and facial keypoints, and produces a high-quality representation of human geometry and view-dependent appearance. The core intuition behind our method is that better drivability and generalization can be achieved by disentangling the driving signals and remaining generative factors, which are not available during animation. To this end, we explicitly account for information deficiency in the driving signal by introducing a latent space that exclusively captures the remaining information, thus enabling the imputation of the missing factors required during full-body animation, while remaining faithful to the driving signal. We also propose a learnable localized compression for the driving signal which promotes better generalization, and helps minimize the influence of global chance-correlations often found in real datasets. For a given driving signal, the resulting variational model produces a compact space of uncertainty for missing factors that allows for an imputation strategy best suited to a particular application. We demonstrate the efficacy of our approach on the challenging problem of full-body animation for virtual telepresence with driving signals acquired from minimal sensors placed in the environment and mounted on a VR-headset.

Deep relightable textures

The increasing demand for 3D content in augmented and virtual reality has motivated the development of volumetric performance capture systemsnsuch as the Light Stage. Recent advances are pushing free viewpoint relightable videos of dynamic human performances closer to photorealistic quality. However, despite significant efforts, these sophisticated systems are limited by reconstruction and rendering algorithms which do not fully model complex 3D structures and higher order light transport effects such as global illumination and sub-surface scattering. In this paper, we propose a system that combines traditional geometric pipelines with a neural rendering scheme to generate photorealistic renderings of dynamic performances under desired viewpoint and lighting. Our system leverages deep neural networks that model the classical rendering process to learn implicit features that represent the view-dependent appearance of the subject independent of the geometry layout, allowing for generalization to unseen subject poses and even novel subject identity. Detailed experiments and comparisons demonstrate the efficacy and versatility of our method to generate high-quality results, significantly outperforming the existing state-of-the-art solutions.

Shape and Reflectance Reconstruction Using Concentric Multi-Spectral Light Field.

Recovering the shape and reflectance of non-Lambertian surfaces remains a challenging problem in computer vision since the view-dependent appearance invalidates traditional photo-consistency constraint. In this paper, we introduce a novel concentric multi-spectral light field (CMSLF) design that is able to recover the shape and reflectance of surfaces of various materials in one shot. Our CMSLF system consists of an array of cameras arranged on concentric circles where each ring captures a specific spectrum. Coupled with a multi-spectral ring light, we are able to sample viewpoint and lighting variations in a single shot via spectral multiplexing. We further show that our concentric camera and light source setting results in a unique single-peak pattern in specularity variations across viewpoints. This property enables robust depth estimation for specular points. To estimate depth and multi-spectral reflectance map, we formulate a physics-based reflectance model for the CMSLF under the surface camera (S-Cam) representation. Extensive synthetic and real experiments show that our method outperforms the state-of-the-art shape reconstruction methods, especially for non-Lambertian surfaces.

Computational Vision for Automatic Tracking and Objective Estimation of Mobile Robot Trajectory

Automatic tracking and evaluation of moving-object trajectories is critical in many applications such as performance estimation of mobile robot navigation. Mobile robot is an effective platform for stimulating student motivation at K-12 institutions as well as a good tool for rigorous engineering practices in colleges, universities, and graduate schools. Developing new mobile robot platforms and algorithms requires objective estimation of navigation performance in a quantitative manner. Conventional methods to estimate mobile robot navigation typically rely on manual usage of chronometer to measure the time spent for the completion of a given task or counting the success rate on the task. This paper proposes an alternative; a multi-camera vision system that can automatically track the movement of mobile robot and estimate it in terms of physics-based profiles: position, velocity, and acceleration of the robot in the trajectory with respect to a user-defined world-coordinate system. The proposed vision system runs two synchronized cameras to simultaneously capture and track the movement of the robot at 30 frames per second. The system runs a homography-based projection algorithm that converts the view-dependent appearance of the robot in the camera images to a view-independent orthographic projection mapped on the registered world coordinate system. This enables the human evaluator to view and estimate the robot navigation from a virtual top-down view embedded with the physics-based profiles regardless of the actual cameras’ viewing positions. The proposed system can also be used for other domains including highway traffic monitoring and intelligent video surveillance.

An adaptive superpixel based hand gesture tracking and recognition system.

We propose an adaptive and robust superpixel based hand gesture tracking system, in which hand gestures drawn in free air are recognized from their motion trajectories. First we employed the motion detection of superpixels and unsupervised image segmentation to detect the moving target hand using the first few frames of the input video sequence. Then the hand appearance model is constructed from its surrounding superpixels. By incorporating the failure recovery and template matching in the tracking process, the target hand is tracked by an adaptive superpixel based tracking algorithm, where the problem of hand deformation, view-dependent appearance invariance, fast motion, and background confusion can be well handled to extract the correct hand motion trajectory. Finally, the hand gesture is recognized by the extracted motion trajectory with a trained SVM classifier. Experimental results show that our proposed system can achieve better performance compared to the existing state-of-the-art methods with the recognition accuracy 99.17% for easy set and 98.57 for hard set.

Bi-scale radiance transfer

Radiance transfer represents how generic source lighting is shadowed and scattered by an object to produce view-dependent appearance. We generalize by rendering transfer at two scales. A macro-scale is coarsely sampled over an object's surface, providing global effects like shadows cast from an arm onto a body. A meso-scale is finely sampled over a small patch to provide local texture. Low-order (25D) spherical harmonics represent low-frequency lighting dependence for both scales. To render, a coefficient vector representing distant source lighting is first transformed at the macro-scale by a matrix at each vertex of a coarse mesh. The resulting vectors represent a spatially-varying hemisphere of lighting incident to the meso-scale. A 4D function, called a radiance transfer texture (RTT), then specifies the surface's meso-scale response to each lighting basis component, as a function of a spatial index and a view direction. Finally, a 25D dot product of the macro-scale result vector with the vector looked up from the RTT performs the correct shading integral. We use an id map to place RTT samples from a small patch over the entire object; only two scalars are specified at high spatial resolution. Results show that bi-scale decomposition makes preprocessing practical and efficiently renders self-shadowing and interreflection effects from dynamic, low-frequency light sources at both scales.

Light field mapping

A light field parameterized on the surface offers a natural and intuitive description of the view-dependent appearance of scenes with complex reflectance properties. To enable the use of surface light fields in real-time rendering we develop a compact representation suitable for an accelerated graphics pipeline. We propose to approximate the light field data by partitioning it over elementary surface primitives and factorizing each part into a small set of lower-dimensional functions. We show that our representation can be further compressed using standard image compression techniques leading to extremely compact data sets that are up to four orders of magnitude smaller than the input data. Finally, we develop an image-based rendering method, light field mapping, that can visualize surface light fields directly from this compact representation at interactive frame rates on a personal computer. We also implement a new method of approximating the light field data that produces positive only factors allowing for faster rendering using simpler graphics hardware than earlier methods. We demonstrate the results for a variety of non-trivial synthetic scenes and physical objects scanned through 3D photography.