Extracting interesting scenarios from real-world data as well as generating failure cases is important for the development and testing of autonomous systems. We propose efficient mechanisms to both characterize and generate testing scenarios using a state-of-the-art driving simulator. For any scenario, our method generates a set of possible driving paths and identifies all the possible safe driving trajectories that can be taken starting at different times, to compute metrics that quantify the complexity of the scenario. We use our method to characterize real driving data from the Next Generation Simulation (NGSIM) project, as well as adversarial scenarios generated in simulation. We rank the scenarios by defining metrics based on the complexity of avoiding accidents and provide insights into how the AV could have minimized the probability of incurring an accident. We demonstrate a strong correlation between the proposed metrics and human intuition.

Stereo-based depth estimation is a cornerstone of computer vision, with state-of-the-art methods delivering accurate results in real time. For several applications such as autonomous navigation, however, it may be useful to trade accuracy for lower latency. We present Bi3D, a method that estimates depth via a series of binary classifications. Rather than testing if objects are at a particular depth $D$, as existing stereo methods do, it classifies them as being closer or farther than $D$. This property offers a powerful mechanism to balance accuracy and latency. Given a strict time budget, Bi3D can detect objects closer than a given distance in as little as a few milliseconds, or estimate depth with arbitrarily coarse quantization, with complexity linear with the number of quantization levels. Bi3D can also use the allotted quantization levels to get continuous depth, but in a specific depth range. For standard stereo (i.e., continuous depth on the whole range), our method is close to or on par with state-of-the-art, finely tuned stereo methods.

We present Extreme View Synthesis, a solution for novel view extrapolation that works even when the number of input images is small---as few as two. In this context, occlusions and depth uncertainty are two of the most pressing issues, and worsen as the degree of extrapolation increases. We follow the traditional paradigm of performing depth-based warping and refinement, with a few key improvements. First, we estimate a depth probability volume, rather than just a single depth value for each pixel of the novel view. This allows us to leverage depth uncertainty in challenging regions, such as depth discontinuities. After using it to get an initial estimate of the novel view, we explicitly combine learned image priors and the depth uncertainty to synthesize a refined image with less artifacts. Our method is the first to show visually pleasing results for baseline magnifications of up to 30x.

Estimation of 3D motion in a dynamic scene from a temporal pair of images is a core task in many scene understanding problems. In real-world applications, a dynamic scene is commonly captured by a moving camera (i.e., panning, tilting or hand-held), increasing the task complexity because the scene is observed from different viewpoints. The primary challenge is the disambiguation of the camera motion from scene motion, which becomes more difficult as the amount of rigidity observed decreases, even with successful estimation of 2D image correspondences. Compared to other state-of-the-art 3D scene flow estimation methods, in this paper, we propose to learn the rigidity of a scene in a supervised manner from an extensive collection of dynamic scene data, and directly infer a rigidity mask from two sequential images with depths. With the learned network, we show how we can effectively estimate camera motion and projected scene flow using computed 2D optical flow and the inferred rigidity mask. For training and testing the rigidity network, we also provide a new semi-synthetic dynamic scene dataset (synthetic foreground objects with a real background) and an evaluation split that accounts for the percentage of observed non-rigid pixels. Through our evaluation, we show the proposed framework outperforms current state-of-the-art scene flow estimation methods in challenging dynamic scenes.

Finding meaningful, structured representations of 3D point cloud data (PCD) has become a core task for spatial perception applications. In this paper we introduce a method for constructing compact generative representations of PCD at multiple levels of detail. As opposed to deterministic structures such as voxel grids or octrees, we propose probabilistic subdivisions of the data through local mixture modeling, and show how these subdivisions can provide a maximum likelihood segmentation of the data. The final representation is hierarchical, compact, parametric, and statistically derived, facilitating run-time occupancy calculations through stochastic sampling. Unlike traditional deterministic spatial subdivision methods, our technique enables dynamic creation of voxel grids according the application's best needs. In contrast to other generative models for PCD, we explicitly enforce sparsity among points and mixtures, a technique which we call expectation sparsification. This leads to a highly parallel hierarchical Expectation Maximization (EM) algorithm well-suited for the GPU and real-time execution. We explore the trade-offs between model fidelity and model size at various levels of detail, our tests showing favorable performance when compared to octree and NDT-based methods.

Registration of Point Cloud Data (PCD) forms a core component of many 3D vision algorithms such as object matching and environment reconstruction. In this paper, we introduce a PCD registration algorithm that utilizes Gaussian Mixture Models (GMM) and a novel dual-mode parameter optimization technique which we call mixture decoupling. We show how this decoupling technique facilitates both faster and more robust registration by first optimizing over the mixture parameters (decoupling the mixture weights, means, and co variances from the points) before optimizing over the 6 DOF registration parameters. Furthermore, we frame both the decoupling and registration process inside a unified, dual-mode Expectation Maximization (EM) framework, for which we derive a Maximum Likelihood Estimation (MLE) solution along with a parallel implementation on the GPU. We evaluate our MLE-based mixture decoupling (MLMD) registration method over both synthetic and real data, showing better convergence for a wider range of initial conditions and higher speeds than previous state of the art methods.

Image registration for stack-based HDR photography is challenging. If not properly accounted for, camera motion and scene changes result in artifacts in the composite image. Unfortunately, existing methods to address this problem are either accurate, but too slow for mobile devices, or fast, but prone to failing. We propose a method that fills this void: our approach is extremely fast - under 700ms on a commercial tablet for a pair of 5MP images - and prevents the artifacts that arise from insufficient registration quality.

We present a method for video stabilization and rolling-shutter correction for videos captured on mobile devices. The method uses the data from an on-board gyroscope to track the camera's angular velocity, and can run in real time within the camera capture pipeline. We remove small motions and rolling-shutter distortions due to hand shake, creating the impression of a video shot on a tripod. For larger motions, we filter the camera's angular velocity to produce a smooth output. To meet the latency constraints of a real-time camera capture pipeline, our filter operates on a small temporal window of three to five frames. Our algorithm performs better than the previous work that uses a gyroscope to stabilize a video stream, and at a similar level with respect to current feature-based methods.

In this chapter we cover the FCam (short for Frankencamera) architecture and API for computational cameras.We begin with the motivation, which is flexible programming of cameras, especially of camera phones and tablets. We cover the API and several example programs that run on the NVIDIA Tegra 3 prototype tablet and the Nokia N900 and N9 Linux-based phones. We discuss the implementation and porting of FCam to different platforms. We also describe how FCam has been used at many universities to teach computational photography.

We propose a new stereo technique using a pair of flash and no-flash stereo images that is both efficient and robust in handling occlusion boundaries. Our work is motivated by the observation that the brightness variations introduced by the flash can provide a robust cue for establishing stereo matches at occlusion boundaries. This photometric cue is computed per pixel, and though on its own is not robust to reliably resolve depth, it can provide a new discriminant to support patch-based stereo matching algorithms. Our experiments using a hand-held Fujifilm W3 3D camera show satisfying stereo performance over a variety of scenes, including several outdoor scenes.

The Frankencamera (FCam) architecture and API enables precise control over the camera in computational photography applications. We present an extension to FCam API for systems equipped with multiple cameras. The proposed extension allows for an enumeration of cameras and their corresponding properties, such as position or orientation. In addition, we explicitly support camera synchronization, either through hardware mechanisms or software primitives. If hardware synchronization is available, cameras can be grouped together under a concept of a multi-sensor. Otherwise, multiple camera streams are scheduled asynchronously and synchronized using our software control primitives.

This course is designed to demonstrate the use of the FCam, an Application Programming Interface (API) which allows for easy and precise control of camera systems. It lets programmers implement advanced computational photography applications. Such applications even allow the modification of basic camera routines such as metering, which are hidden inside black boxes in most camera systems. FCam is also an excellent tool for teaching and research on interactive mobile computational photography. The API has been open-sourced, and is currently implemented for Nokia N900 and an experimental Frankencamera (the F2). The Frankencamera architecture for computational cameras was presented at SIGGRAPH 2010. By end-2011, the FCam will hopefully be available on Nokia™ Linux phones, NVIDIA™ Tegra 3 development boards, and F3, the next version of the Stanford experimental camera. The F3 will be distributed to research universities within the US along with courseware in a National Science Foundation (NSF)-sponsored project. NVIDIA Tegras and Nokia™ Linux phones are available globally.

Systems for the creation of photorealistic models using range scans and digital photographs are becoming increasingly popular in a wide range of fields, from reverse engineering to cultural heritage preservation. These systems employ a range finder to acquire the geometry information and a digital camera to measure color detail. But bringing together a set of range scans and color images to produce an accurate and usable model is still an area of research with many unsolved problems. In this paper we address the problem of how to build illumination coherent integrated texture maps from images that were taken under different illumination conditions. To achieve this we present two different solutions. The first one is to align all the images to the same illumination, for which we have developed a technique that computes a relighting operator over the area of overlap of a pair of images that we then use to relight the entire image. Our proposed method can handle images with shadows and can effectively remove the shadows from the image, if required. The second technique uses the ratio of two images to factor out the diffuse reflectance of an image from its illumination. We do this without any light measuring device. By computing the actual reflectance we remove from the images any effects of the illumination, allowing us to create new renderings under novel illumination conditions.

This paper presents a novel method for the registration of texture images with a 3D model of outdoor scenes. We pose image registration as an optimization problem that uses knowledge of the sun’s position to estimate shadows in a scene, and use the shadows produced as a cue to solve for the registration parameters. Results are presented on a controlled experiment and for a large scale model of an archaeological site in Sicily.

In this paper we consider the problem of factoring illumination and texture from a pair of images of a diffuse object of known geometry. This problem arises frequently in 3D photography applications that use images to acquire photometric properties of a scanned object. Our approach uses the ratio of the images and the geometry information to compute the relative incident irradiance of one image with respect to the other. After the irradiance maps are recovered, we build a spatially varying albedo map, which can then be used to render the object under different illumination conditions. We present two algorithms, one for point-light source illumination, and another one based on spherical harmonics for more general illumination conditions.

Systems for the creation of photorealistic models using range scans and digital photographs are becoming increasingly popular in a wide range of fields, from reverse engineering to cultural heritage preservation. These systems employ a range finder to acquire the geometry information and a digital camera to measure color detail. But bringing together a set of range scans and color images to produce an accurate and usable model is still an area of research with many unsolved problems. In this dissertation we present new tools and methods for creating digital models from range and color images, with emphasis in large-scale outdoor scenes. First, we address the problem of range and color image registration. In this area, we introduce a semi-automatic tool for range and color image registration that makes use of line-features to solve for the position and orientation of the digital camera. This allows us to efficiently register images of urban scenery. Secondly, we present a registration technique that uses the shadows cast by the sun as cues find the correct camera pose, which we have successfully applied in the creation of a digital model of an archaeological excavation in Monte Polizzo, Sicily. We also address the problem of how to build seamless integrated texture maps from images that were taken under different illumination conditions. To achieve this we present two different solutions. The first one is to align all the images to the same illumination. For this, we have developed a technique that computes a relighting operator over the area of overlap of a pair of images, which we then use to relight the entire image. Our proposed method can handle images with shadows and can effectively remove the shadows from the image; if required. The second technique uses the ratio of two images to factor out the diffuse reflectance of an image from its illumination. We achieve this without any light measuring device. By computing the actual reflectance we remove from the images any effects of the illumination, which then allows us to create new renderings under novel illumination conditions.