Search

Recently, several approaches have emerged for generating neural representations with multiple levels of detail (LODs). LODs can improve the rendering by using lower resolutions and smaller model sizes when appropriate. However, existing methods generally focus on a few discrete LODs which suffer from aliasing and flicker artifacts as details are changed and limit their granularity for adapting to resource limitations. In this paper, we propose a method to encode light field networks with continuous LODs, allowing for finely tuned adaptations to rendering conditions. Our training procedure uses summed-area table filtering allowing efficient and continuous filtering at various LODs. Furthermore, we use saliency-based importance sampling which enables our light field networks to distribute their capacity, particularly limited at lower LODs, towards representing the details viewers are most likely to focus on. Incorporating continuous LODs into neural representations enables progressive streaming of neural representations, decreasing the latency and resource utilization for rendering., Comment: Accepted to BMVC 2023. Webpage at https://augmentariumlab.github.io/continuous-lfn/

We present a novel technique for implicit neural representation of light fields at continuously defined viewpoints with high quality and fidelity. Our implicit neural representation maps 4D coordinates defining two-plane parameterization of the light fields to the corresponding color values. We leverage periodic activations to achieve high expressivity and accurate reconstruction for complex data manifolds while keeping low storage and inference time requirements. However, na\"ively trained non-3D structured networks do not adequately satisfy the multi-view consistency; instead, they perform alpha blending of nearby viewpoints. In contrast, our View Correspondence Network, or VICON, leverages stereo matching, optimization by automatic differentiation with respect to the input space, and multi-view pixel correspondence to provide a novel implicit representation of the light fields faithful to the novel views that are unseen during the training. Experimental results show VICON superior to the state-of-the-art non-3D implicit light field representations both qualitatively and quantitatively. Moreover, our implicit representation captures a larger field of view (FoV), surpassing the extent of the observable scene by the cameras of the ground truth renderings., Comment: 10 pages, 7 figures

We present VIINTER, a method for view interpolation by interpolating the implicit neural representation (INR) of the captured images. We leverage the learned code vector associated with each image and interpolate between these codes to achieve viewpoint transitions. We propose several techniques that significantly enhance the interpolation quality. VIINTER signifies a new way to achieve view interpolation without constructing 3D structure, estimating camera poses, or computing pixel correspondence. We validate the effectiveness of VIINTER on several multi-view scenes with different types of camera layout and scene composition. As the development of INR of images (as opposed to surface or volume) has centered around tasks like image fitting and super-resolution, with VIINTER, we show its capability for view interpolation and offer a promising outlook on using INR for image manipulation tasks., Comment: SIGGRAPH Asia 2022

Neural representations have shown great promise in their ability to represent radiance and light fields while being very compact compared to the image set representation. However, current representations are not well suited for streaming as decoding can only be done at a single level of detail and requires downloading the entire neural network model. Furthermore, high-resolution light field networks can exhibit flickering and aliasing as neural networks are sampled without appropriate filtering. To resolve these issues, we present a progressive multi-scale light field network that encodes a light field with multiple levels of detail. Lower levels of detail are encoded using fewer neural network weights enabling progressive streaming and reducing rendering time. Our progressive multi-scale light field network addresses aliasing by encoding smaller anti-aliased representations at its lower levels of detail. Additionally, per-pixel level of detail enables our representation to support dithered transitions and foveated rendering., Comment: Accepted to 3DV 2022. Website at https://augmentariumlab.github.io/multiscale-lfn/

We introduce a new implicit shape representation called Primary Ray-based Implicit Function (PRIF). In contrast to most existing approaches based on the signed distance function (SDF) which handles spatial locations, our representation operates on oriented rays. Specifically, PRIF is formulated to directly produce the surface hit point of a given input ray, without the expensive sphere-tracing operations, hence enabling efficient shape extraction and differentiable rendering. We demonstrate that neural networks trained to encode PRIF achieve successes in various tasks including single shape representation, category-wise shape generation, shape completion from sparse or noisy observations, inverse rendering for camera pose estimation, and neural rendering with color., Comment: ECCV 2022. Project Page: https://augmentariumlab.github.io/PRIF/

With the recent interest in virtual reality and augmented reality, there is a newfound demand for displays that can provide high resolution with a wide field of view (FOV). However, such displays incur significantly higher costs for rendering the larger number of pixels. This poses the challenge of rendering realistic real-time images that have a wide FOV and high resolution using limited computing resources. The human visual system does not need every pixel to be rendered at a uniformly high quality. Foveated rendering methods provide perceptually high-quality images while reducing computational workload and are becoming a crucial component for large-scale rendering. In this paper, we present key motivations, research directions, and challenges for leveraging the limitations of the human visual system as they relate to foveated rendering. We provide a taxonomy to compare and contrast various foveated techniques based on key factors. We also review aliasing artifacts arising due to foveation methods and discuss several approaches that attempt to mitigate such effects. Finally, we present several open problems and possible future research directions that can further reduce computational costs while generating perceptually high-quality renderings., Comment: 16 pages, 12 figures

Immersive maps such as Google Street View and Bing Streetside provide true-to-life views with a massive collection of panoramas. However, these panoramas are only available at sparse intervals along the path they are taken, resulting in visual discontinuities during navigation. Prior art in view synthesis is usually built upon a set of perspective images, a pair of stereoscopic images, or a monocular image, but barely examines wide-baseline panoramas, which are widely adopted in commercial platforms to optimize bandwidth and storage usage. In this paper, we leverage the unique characteristics of wide-baseline panoramas and present OmniSyn, a novel pipeline for 360{\deg} view synthesis between wide-baseline panoramas. OmniSyn predicts omnidirectional depth maps using a spherical cost volume and a monocular skip connection, renders meshes in 360{\deg} images, and synthesizes intermediate views with a fusion network. We demonstrate the effectiveness of OmniSyn via comprehensive experimental results including comparison with the state-of-the-art methods on CARLA and Matterport datasets, ablation studies, and generalization studies on street views. We envision our work may inspire future research for this unheeded real-world task and eventually produce a smoother experience for navigating immersive maps., Comment: Updated related works

Federated learning (FL) is a popular paradigm for private and collaborative model training on the edge. In centralized FL, the parameters of a global architecture (such as a deep neural network) are maintained and distributed by a central server/controller to clients who transmit model updates (gradients) back to the server based on local optimization. While many efforts have focused on reducing the communication complexity of gradient transmission, the vast majority of compression-based algorithms assume that each participating client is able to download and train the current and full set of parameters, which may not be a practical assumption depending on the resource constraints of smaller clients such as mobile devices. In this work, we propose a simple yet effective novel algorithm, Comfetch, which allows clients to train large networks using reduced representations of the global architecture via the count sketch, which reduces local computational and memory costs along with bi-directional communication complexity. We provide a nonconvex convergence guarantee and experimentally demonstrate that it is possible to learn large models, such as a deep convolutional network, through federated training on their sketched counterparts. The resulting global models exhibit competitive test accuracy over CIFAR10/100 classification when compared against un-compressed model training.