Intersection management systems, with the assistance of vehicular networks and autonomous vehicles, have the potential to perform traffic control more precisely than contemporary signalized intersections. However, as infrastructural intersection management controllers do not directly activate motions of vehicles, it is possible that the vehicles fail to follow the instructions from controllers, undermining system properties such as deadlock-freeness and traffic performance. In this article, we consider a class of robustness issues, the time violations, which stem from possible discrepancies between scheduled orders and real executions. We refine a graph-based intersection model to build our theoretical foundations and analyze potential deadlocks and their resolvability. We develop solutions that mitigate negative effects of time violations. In particular, we propose a Robustness-Aware Greedy Scheduling algorithm for robust scheduling and evaluate the deadlock-free robustness of different intersection models and scheduling algorithms. Experimental results show that the Robustness-Aware Greedy Scheduling algorithm is able to significantly improve robustness and keep a good balance with traffic performance.

The goal of intersection management is to organize vehicles to pass the intersection safely and efficiently. Due to the technical advance of connected and autonomous vehicles, intersection management becomes more intelligent and potentially unsignalized. In this paper, we propose a reinforcement-learning-based methodology to train a centralized intersection manager. We define the intersection scheduling problem with a graph-based model and transform it to the job-shop scheduling problem (JSSP) with additional constraints. To utilize reinforcement learning, we model the scheduling procedure as a Markov decision process (MDP) and train the agent with the proximal policy optimization (PPO). A grouping strategy is also developed to apply the trained model to streams of vehicles. Experimental results show that the learning-based intersection manager is especially effective with high traffic densities. This paper is the first work in the literature to apply reinforcement learning on the graph-based intersection model. The proposed methodology can flexibly deal with any conflicting scenario and indicate the applicability of reinforcement learning to Intelligent intersection management.

Novel view synthesis from a single image requires inferring occluded regions of objects and scenes whilst simultaneously maintaining semantic and physical consistency with the input. Existing approaches condition neural radiance fields (NeRF) on local image features, projecting points to the input image plane, and aggregating 2D features to perform volume rendering. However, under severe occlusion, this projection fails to resolve uncertainty, resulting in blurry renderings that lack details. In this work, we propose NerfDiff, which addresses this issue by distilling the knowledge of a 3D-aware conditional diffusion model (CDM) into NeRF through synthesizing and refining a set of virtual views at test-time. We further propose a novel NeRF-guided distillation algorithm that simultaneously generates 3D consistent virtual views from the CDM samples, and finetunes the NeRF based on the improved virtual views. Our approach significantly outperforms existing NeRF-based and geometry-free approaches on challenging datasets including ShapeNet, ABO, and Clevr3D.

Although neural radiance fields (NeRF) have shown impressive advances in novel view synthesis, most methods require multiple input images of the same scene with accurate camera poses. In this work, we seek to substantially reduce the inputs to a single unposed image. Existing approaches using local image features to reconstruct a 3D object often render blurry predictions at viewpoints distant from the source view. To address this, we propose to leverage both the global and local features to form an expressive 3D representation. The global features are learned from a vision transformer, while the local features are extracted from a 2D convolutional network. To synthesize a novel view, we train a multi-layer perceptron (MLP) network conditioned on the learned 3D representation to perform volume rendering. This novel 3D representation allows the network to reconstruct unseen regions without enforcing constraints like symmetry or canonical coordinate systems. Our method renders novel views from just a single input image, and generalizes across multiple object categories using a single model. Quantitative and qualitative evaluations demonstrate that the proposed method achieves state-of-the-art performance and renders richer details than existing approaches. https://cseweb.ucsd.edu/ %7eviscomp/projects/VisionNeRF/

Abstract Portrait synthesis creates realistic digital avatars which enable users to interact with others in a compelling way. Recent advances in StyleGAN and its extensions have shown promising results in synthesizing photorealistic and accurate reconstruction of human faces. However, previous methods often focus on frontal face synthesis and most methods are not able to handle large head rotations due to the training data distribution of StyleGAN. In this work, our goal is to take as input a monocular video of a face, and create an editable dynamic portrait able to handle extreme head poses. The user can create novel viewpoints, edit the appearance, and animate the face. Our method utilizes pivotal tuning inversion (PTI) to learn a personalized video prior from a monocular video sequence. Then we can input pose and expression coefficients to MLPs and manipulate the latent vectors to synthesize different viewpoints and expressions of the subject. We also propose novel loss functions to further disentangle pose and expression in the latent space. Our algorithm shows much better performance over previous approaches on monocular video datasets, and it is also capable of running in real‐time at 54 FPS on an RTX 3080.

Abstract Precomputed Radiance Transfer (PRT) remains an attractive solution for real‐time rendering of complex light transport effects such as glossy global illumination. After precomputation, we can relight the scene with new environment maps while changing viewpoint in real‐time. However, practical PRT methods are usually limited to low‐frequency spherical harmonic lighting. All‐frequency techniques using wavelets are promising but have so far had little practical impact. The curse of dimensionality and much higher data requirements have typically limited them to relighting with fixed view or only direct lighting with triple product integrals. In this paper, we demonstrate a hybrid neural‐wavelet PRT solution to high‐frequency indirect illumination, including glossy reflection, for relighting with changing view. Specifically, we seek to represent the light transport function in the Haar wavelet basis. For global illumination, we learn the wavelet transport using a small multi‐layer perceptron (MLP) applied to a feature field as a function of spatial location and wavelet index, with reflected direction and material parameters being other MLP inputs. We optimize/learn the feature field (compactly represented by a tensor decomposition) and MLP parameters from multiple images of the scene under different lighting and viewing conditions. We demonstrate real‐time (512 x 512 at 24 FPS, 800 x 600 at 13 FPS) precomputed rendering of challenging scenes involving view‐dependent reflections and even caustics.

Image view synthesis has seen great success in reconstructing photorealistic visuals, thanks to deep learning and various novel representations. The next key step in immersive virtual experiences is view synthesis of dynamic scenes. However, several challenges exist due to the lack of high-quality training datasets, and the additional time dimension for videos of dynamic scenes. To address this issue, we introduce a multi-view video dataset, captured with a custom 10-camera rig in 120FPS. The dataset contains 96 high-quality scenes showing various visual effects and human interactions in outdoor scenes. We develop a new algorithm, Deep 3D Mask Volume, which enables temporally-stable view extrapolation from binocular videos of dynamic scenes, captured by static cameras. Our algorithm addresses the temporal inconsistency of disocclusions by identifying the error-prone areas with a 3D mask volume, and replaces them with static background observed throughout the video. Our method enables manipulation in 3D space as opposed to simple 2D masks, We demonstrate better temporal stability than frame-by-frame static view synthesis methods, or those that use 2D masks. The resulting view synthesis videos show minimal flickering artifacts and allow for larger translational movements.

Intelligent intersection management aims to schedule vehicles so that vehicles can pass through an intersection efficiently and safely. However, inaccurate control, imperfect communication, and malicious information or behavior lead to robustness issues of intelligent intersection management. In this work, we focus on improving robustness against deadlocks by changing the trajectories of vehicles. To guarantee the resolvability of deadlocks, we limit the number of vehicles in an intersection to be smaller than or equal to an intersection-specific value called the maximal deadlock-free load. We develop an algorithm to compute the maximal deadlock-free load. We further reduce the computation time by computing the loads which are pessimistic (smaller) but still deadlock-free. Since the maximal deadlock-free load only depends on the given intersection, it can be integrated with different scheduling algorithms. Experimental results demonstrate that, by changing the trajectories of vehicles and limiting the number of vehicles under maximal deadlock-free loads, our approach can guarantee deadlock-freeness and maintain good traffic efficiency.

Image view synthesis has seen great success in reconstructing photorealistic visuals, thanks to deep learning and various novel representations. The next key step in immersive virtual experiences is view synthesis of dynamic scenes. However, several challenges exist due to the lack of high-quality training datasets, and the additional time dimension for videos of dynamic scenes. To address this issue, we introduce a multi-view video dataset, captured with a custom 10-camera rig in 120FPS. The dataset contains 96 high-quality scenes showing various visual effects and human interactions in outdoor scenes. We develop a new algorithm, Deep 3D Mask Volume, which enables temporally-stable view extrapolation from binocular videos of dynamic scenes, captured by static cameras. Our algorithm addresses the temporal inconsistency of disocclusions by identifying the error-prone areas with a 3D mask volume, and replaces them with static background observed throughout the video. Our method enables manipulation in 3D space as opposed to simple 2D masks, We demonstrate better temporal stability than frame-by-frame static view synthesis methods, or those that use 2D masks. The resulting view synthesis videos show minimal flickering artifacts and allow for larger translational movements.

Human perception of the visual world in the presence of fog or haze usually suffers from loss of color vibrance and details. In this paper we propose a novel method to provide a see-through head-mounted display with image dehazing capability. The proposed method projects an auxiliary image from the head-mounted device to the user's retina. The introduction of the auxiliary image enhances the contrast and brightness of the perceived visual world. This method can be applied to tourism, military, and navigation to overcome poor visual condition.