Projects

When capturing a light field of a scene, one typically faces a trade-off between more spatial or more angular resolution. Fortunately, light fields are also a rich source of information for solving the problem of super-resolution. Contrary to single image approaches, where high-frequency content has to be hallucinated to be the most likely source of the downscaled version, sub-aperture views from the light field can help with an actual reconstruction of those details that have been removed by downsampling. In this paper, we propose a three-dimensional generative adversarial autoencoder network to recover the high-resolution light field from a low-resolution light field with a sparse set of viewpoints. We require only three views along both horizontal and vertical axis to increase angular resolution by a factor of three while at the same time increasing spatial resolution by a factor of either two or four in each direction, respectively.

We propose a novel idea to introduce regularization based on second order total generalized variation (TGV) into optimization frameworks based on functional lifting. The proposed formulation extends a recent sublabel-accurate relaxation for multi-label problems and thus allows for accurate solutions using only a small number of labels, significantly improving over previous approaches towards lifting the total generalized variation. Moreover, even recent sublabel accurate methods exhibit staircasing artifacts when used in conjunction with common first order regularizers such as the total variation (TV). This becomes very obvious for example when computing derivatives of disparity maps computed with these methods to obtain normals, which immediately reveals their local flatness and yields inaccurate normal maps. We show that our approach is effective in reducing these artifacts, obtaining disparity maps with a smooth normal field in a single optimization pass.

We present an encoder-decoder deep neural network that solves non-Lambertian intrinsic light field decomposition, where we recover all three intrinsic components: albedo, shading, and specularity. We learn a sparse set of features from 3D epipolar volumes and use them in separate decoder pathways to reconstruct intrinsic light fields. While being trained on synthetic data generated with Blender, our model still generalizes to real world examples captured with a Lytro Illum plenoptic camera. The proposed method outperforms state-of-the-art approaches for single images and achieves competitive accuracy with recent modeling methods for light fields.

We present a real-time pipeline for large-scale 3D scene reconstruction from a single moving RGB-D camera together with interactive visualization. Our approach combines a time and space efficient data structure capable of representing large scenes, a local variational update algorithm and a visualization system. The environment's structure is reconstructed by integrating the depth image of each camera view into a sparse volume representation using a truncated signed distance function, which is organized via a hash table. Noise from real-world data is efficiently eliminated by immediately performing local variational refinements on newly integrated data. The whole pipeline is able to perform in real-time on consumer-available hardware and allows for simultaneous inspection of the currently reconstructed scene.

In this work, we focus on the problem of intrinsic scene decomposition in light fields. Our main contribution is a novel prior to cope with cast shadows and inter-reflections. In contrast to other approaches which model inter-reflection based only on geometry, we model indirect shading by combining geometric and color information. We compute a shadow confidence measure for the light field and use it in the regularization constraints. Another contribution is an improved specularity estimation by using color information from sub-aperture views. The new priors are embedded in a recent framework to decompose the input light field into albedo, shading, and specularity. We arrive at a variational model where we regularize albedo and the two shading components on epipolar plane images, encouraging them to be consistent across all sub-aperture views. Our method is evaluated on ground truth synthetic datasets and real world light fields. We outperform both state-of-the art approaches for RGB+D images and recent methods proposed for light fields.

We propose an inverse rendering model for light fields to recover surface normals, depth, reflectance and natural illumination. Our setting is fully uncalibrated, with the reflectance modeled with a spatially-constant Blinn-Phong model and illumination as an environment map. While previous work makes strong assumptions in this difficult scenario, focusing solely on specific types of objects like faces or imposing very strong priors, our approach leverages only the light field structure, where a solution consistent across all subaperture views is sought. The optimization is based primarily on shading, which is sensitive to fine geometric details which are propagated to the initial coarse depth map. Despite the problem being inherently ill-posed, we achieve encouraging results on synthetic as well as real-world data.

We present a fully convolutional autoencoder for light fields, which jointly encodes stacks of horizontal and vertical epipolar plane images through a deep network of residual layers. The complex structure of the light field is thus reduced to a comparatively low-dimensional representation, which can be decoded in a variety of ways. The different pathways of upconvolution we currently support are for disparity estimation and separation of the lightfield into diffuse and specular intrinsic components. The key idea is that we can jointly perform unsupervised training for the autoencoder path of the network, and supervised training for the other decoders. This way, we find features which are both tailored to the respective tasks and generalize well to datasets for which only example light fields are available. We provide an extensive evaluation on synthetic light field data, and show that the network yields good results on previously unseen real world data captured by a Lytro Illum camera and various gantries.

This paper presents the results of the depth estimation challenge for dense light fields, which took place at the second workshop on Light Fields for Computer Vision (LF4CV) in conjunction with CVPR 2017. The challenge consisted of submission to a recent benchmark [7], which allows a thorough performance analysis. While individual results are readily available on the benchmark web page www.lightfield-analysis.net, we take this opportunity to give a detailed overview of the current participants. Based on the algorithms submitted to our challenge, we develop a taxonomy of light field disparity estimation algorithms and give a report on the current state-of-the-art. In addition, we include more comparative metrics, and discuss the relative strengths and weaknesses of the algorithms. Thus, we obtain a snapshot of where light field algorithm development stands at the moment and identify aspects with potential for further improvement.

We introduce a novel approach to jointly estimate consistent depth and normal maps from 4D light fields, with two main contributions. First, we build a cost volume from focal stack symmetry. However, in contrast to previous approaches, we introduce partial focal stacks in order to be able to robustly deal with occlusions. This idea already yields significanly better disparity maps. Second, even recent sublabel-accurate methods for multi-label optimization recover only a piecewise flat disparity map from the cost volume, with normals pointing mostly towards the image plane. This renders normal maps recovered from these approaches unsuitable for potential subsequent applications. We therefore propose regularization with a novel prior linking depth to normals, and imposing smoothness of the resulting normal field. We then jointly optimize over depth and normals to achieve estimates for both which surpass previous work in accuracy on a recent benchmark.

We present a reconstruction pipeline for a large-scale 3D environment viewed by a single moving RGB-D camera. Our approach combines advantages of fast and direct, regularization-free depth fusion and accurate, but costly variational schemes. The scene’s depth geometry is extracted from each camera view and efficiently integrated into a large, dense grid as a truncated signed distance function, which is organized in an octree. To account for noisy real-world input data, variational range image integration is performed in local regions of the volume directly on this octree structure. We focus on algorithms which are easily parallelizable on GPUs, allowing the pipeline to be used in real-time scenarios where the user can interactively view the reconstruction and adapt camera motion as required.

In this paper, we propose a novel method for depth estimation in light fields which employs a specifically designed sparse decomposition to leverage the depth-orientation relationship on its epipolar plane images. The proposed method learns the structure of the central view and uses this information to construct a light field dictionary for which groups of atoms correspond to unique disparities. This dictionary is then used to code a sparse representation of the light field. Analyzing the coefficients of this representation with respect to the disparities of their corresponding atoms yields an accurate and robust estimate of depth. In addition, if the light field has multiple depth layers, such as for reflective or transparent surfaces, statistical analysis of the coefficients can be employed to infer the respective depth of the superimposed layers.

We present a method to separate a dichromatic reflection component from diffuse object colors for the set of rays in a 4D light field such that the separation is consistent across all subaperture views. The separation model is based on explaining the observed light field as a sparse linear combination of a constant-color specular term and a small finite set of albedos. Consistency across the light field is achieved by embedding the ray-wise separation into a global optimization framework. On each individual epipolar plane image (EPI), the diffuse coefficients need to be constant along lines which are the projections of the same scene point, while the specular coefficient needs to be constant along the direction of the specular flow within the epipolar volume. We handle both constraints with depth-dependent anisotropic regularizers, and demonstrate promising performance on a number of real-world light fields captured with a Lytro Illum plenoptic camera.