We present a method for generating video sequences with coherent motion between a pair of input keyframes. We adapt a pretrained large-scale image-to-video diffusion model (originally trained to generate videos moving forward in time from a single input image) for keyframe interpolation, i.e., to produce a video between two input frames. We accomplish this adaptation through a lightweight fine-tuning technique that produces a version of the model that instead predicts videos moving backwards in time from a single input image. This model (along with the original forward-moving model) is subsequently used in a dual-directional diffusion sampling process that combines the overlapping model estimates starting from each of the two keyframes. Our experiments shows that our method outperforms both existing diffusion-based methods and traditional frame interpolation techniques.

Modern GPUs are equipped with tensor cores (TCs) that are commonly used for matrix multiplication in artificial intelligence workloads. However, because they have high computational throughput, they can lead to significant performance gains in other algorithms if they can be successfully exploited. We examine using TCs to compute Euclidean distance calculations, which are used in many data analytics applications. Prior work has only investigated using 64 bit floating point (FP64) data for computation; however, TCs can operate on lower precision floating point data (i.e., 16 bit matrix multiplication and 32 bit accumulation), which we refer to as FP16-32. FP16-32 TC peak throughput is so high that TCs are easily starved of data. We propose a Fast and Scalable Tensor core Euclidean Distance (FaSTED) algorithm. To achieve high computational throughput, we design FaSTED for significant hierarchical reuse of data and maximize memory utilization at every level (global memory, shared memory, and registers). We apply FaSTED to the application of similarity searches, which typically employ an indexing data structure to eliminate superfluous Euclidean distance calculations. We compare to the state-of-the-art (SOTA) TC Euclidean distance algorithm in the literature that employs FP64, as well as to two single precision (FP32) CUDA core algorithms that both employ an index. We find that across four real-world high-dimensional datasets spanning 128-960 dimensions, the mixed-precision brute force approach achieves a speedup over the SOTA algorithms of 2.5–51 ×. We also quantify the accuracy loss of our mixed precision algorithm to be < 0.06% when compared to the FP64 baseline.

We present UltraZoom, a system for generating gigapixel-resolution images of objects from casually captured inputs, such as handheld phone photos. Given a full-shot image (global, low-detail) and one or more close-ups (local, high-detail), UltraZoom upscales the full image to match the fine detail and scale of the close-up examples. To achieve this, we construct a per-instance paired dataset from the close-ups and adapt a pretrained generative model to learn object-specific low-to-high resolution mappings. At inference, we apply the model in a sliding window fashion over the full image. Constructing these pairs is non-trivial: it requires registering the close-ups within the full image for scale estimation and degradation alignment. We introduce a simple, robust method for achieving registration on arbitrary materials in casual, in-the-wild captures. Together, these components form a system that enables seamless pan and zoom across the entire object, producing consistent, photorealistic gigapixel imagery from minimal input. For full-resolution results and code, visit our project page at ultra-zoom.github.io .

We introduce a fully automatic pipeline for dynamic scene reconstruction from casually captured monocular RGB videos. Rather than designing a new scene representation, we enhance the priors that drive Dynamic Gaussian Splatting. Video segmentation combined with epipolar-error maps yields object-level masks that closely follow thin structures; these masks (i) guide an object-depth loss that sharpens the consistent video depth, and (ii) support skeleton-based sampling plus mask-guided re-identification to produce reliable, comprehensive 2-D tracks. Two additional objectives embed the refined priors in the reconstruction stage: a virtual-view depth loss removes floaters, and a scaffold-projection loss ties motion nodes to the tracks, preserving fine geometry and coherent motion. The resulting system surpasses previous monocular dynamic scene reconstruction methods and delivers visibly superior renderings. Project page: https://priorenhancedgaussian.github.io/

We present a method to generate full-body selfies from photographs originally taken at arms length. Because self-captured photos are typically taken close up, they have lim-ited field of view and exaggerated perspective that distorts facial shapes. We instead seek to generate the photo some one else would take of you from a few feet away. Our approach takes as input four selfies of your face and body, a background image, and generates a full-body selfie in a de-sired target pose. We introduce a novel diffusion-based approach to combine all of this information into high-quality, well-composed photos of you with the desired pose and background.

We propose ExtraNeRF, a novel method for extrapolating the range of views handled by a Neural Radiance Field (NeRF). Our main idea is to leverage NeRFs to model scene-specific, fine-grained details, while capitalizing on diffusion models to extrapolate beyond our observed data. A key ingredient is to track visibility to determine what portions of the scene have not been observed, and focus on re-constructing those regions consistently with diffusion mod-els. Our primary contributions include a visibility-aware diffusion-based inpainting module that is fine-tuned on the input imagery, yielding an initial NeRF with moderate qual-ity (often blurry) inpainted regions, followed by a second diffusion model trained on the input imagery to consistently enhance, notably sharpen, the inpainted imagery from the first pass. We demonstrate high-quality results, extrapolating beyond a small number of (typically six or fewer) input views, effectively outpainting the NeRF as well as inpainting newly disoccluded regions inside the original viewing volume. We compare with related work both quantitatively and qualitatively and show significant gains over prior art.

We present a method that uses a text-to-image model to generate consistent content across multiple image scales, enabling extreme semantic zooms into a scene, e.g. ranging from a wide-angle landscape view of a forest to a macro shot of an insect sitting on one of the tree branches. We achieve this through a joint multi-scale diffusion sampling approach that encourages consistency across different scales while preserving the integrity of each individual sampling process. Since each generated scale is guided by a different text prompt, our method enables deeper levels of zoom than traditional super-resolution methods that may struggle to create new contextual structure at vastly different scales. We compare our method qualitatively with alter-native techniques in image super-resolution and outpainting, and show that our method is most effective at generating consistent multi-scale content.

We introduce light diffusion, a novel method to improve lighting in portraits, softening harsh shadows and specular highlights while preserving overall scene illumi-nation. Inspired by professional photographers' diffusers and scrims, our method softens lighting given only a single portrait photo. Previous portrait relighting approaches focus on changing the entire lighting environment, removing shadows (ignoring strong specular highlights), or removing shading entirely. In contrast, we propose a learning based method that allows us to control the amount of light diffusion and apply it on in-the-wild portraits. Additionally, we design a method to synthetically generate plausible external shadows with sub-surface scattering effects while conforming to the shape of the subject's face. Finally, we show how our approach can increase the robustness of higher level vision applications, such as albedo estimation, geometry estimation and semantic segmentation.

We present PersonNeRF, a method that takes a collection of photos of a subject (e.g. Roger Federer) captured across multiple years with arbitrary body poses and ap-pearances, and enables rendering the subject with arbitrary novel combinations of viewpoint, body pose, and appearance. PersonNeRF builds a customized neural volumetric 3D model of the subject that is able to render an entire space spanned by camera viewpoint, body pose, and appearance. A central challenge in this task is dealing with sparse observations; a given body pose is likely only observed by a single viewpoint with a single appearance, and a given appearance is only observed under a handful of different body poses. We address this issue by recovering a canonical T-pose neural volumetric representation of the subject that allows for changing appearance across different observations, but uses a shared pose-dependent motion field across all observations. We demonstrate that this approach, along with regularization of the recovered volumetric geometry to encourage smoothness, is able to recover a model that renders compelling images from novel combinations of viewpoint, pose, and appearance from these challenging unstructured photo collections, outperforming prior work for free-viewpoint human rendering.

We present a system that automatically brings street view imagery to life by populating it with naturally behaving, animated pedestrians and vehicles. Our approach is to remove existing people and vehicles from the input image, insert moving objects with proper scale, angle, motion and appearance, plan paths and traffic behavior, as well as render the scene with plausible occlusion and shadowing effects. The system achieves these by reconstructing the still image street scene, simulating crowd behavior, and rendering with consistent lighting, visibility, occlusions, and shadows. We demonstrate results on a diverse range of street scenes including regular still images and panoramas.

We introduce 3D Moments, a new computational photography effect. As input we take a pair of near-duplicate photos, i.e., photos of moving subjects from similar viewpoints, common in people's photo collections. As output, we produce a video that smoothly interpolates the scene motion from the first photo to the second, while also producing camera motion with parallax that gives a heightened sense of 3D. To achieve this effect, we represent the scene as a pair of feature-based layered depth images augmented with scene flow. This representation enables motion interpolation along with independent control of the camera viewpoint. Our system produces photorealistic space-time videos with motion parallax and scene dynamics, while plausibly recovering regions occluded in the original views. We conduct extensive experiments demonstrating superior performance over baselines on public datasets and in-the-wild photos. Project page: https://3d-moments.github.io/.

We introduce a free-viewpoint rendering method - HumanNeRF - that works on a given monocular video of a human performing complex body motions, e.g. a video from YouTube. Our method enables pausing the video at any frame and rendering the subject from arbitrary new camera viewpoints or even a full 360-degree camera path for that particular frame and body pose. This task is particularly challenging, as it requires synthesizing photorealistic details of the body, as seen from various camera angles that may not exist in the input video, as well as synthesizing fine details such as cloth folds and facial appearance. Our method optimizes for a volumetric representation of the person in a canonical T-pose, in concert with a motion field that maps the estimated canonical representation to every frame of the video via backward warps. The motion field is decomposed into skeletal rigid and non-rigid motions, produced by deep networks. We show significant performance improvements over prior work, and compelling examples of free-viewpoint renderings from monocular video of moving humans in challenging uncontrolled capture scenarios.

In this paper, we demonstrate a fully automatic method for converting a still image into a realistic animated looping video. We target scenes with continuous fluid motion, such as flowing water and billowing smoke. Our method relies on the observation that this type of natural motion can be convincingly reproduced from a static Eulerian motion description, i.e. a single, temporally constant flow field that defines the immediate motion of a particle at a given 2D location. We use an image-to-image translation network to encode motion priors of natural scenes collected from on-line videos, so that for a new photo, we can synthesize a corresponding motion field. The image is then animated using the generated motion through a deep warping technique: pixels are encoded as deep features, those features are warped via Eulerian motion, and the resulting warped feature maps are decoded as images. In order to produce continuous, seamlessly looping video textures, we propose a novel video looping technique that flows features both for-ward and backward in time and then blends the results. We demonstrate the effectiveness and robustness of our method by applying it to a large collection of examples including beaches, waterfalls, and flowing rivers.

We introduce a real-time, high-resolution background replacement technique which operates at 30fps in 4K resolution, and 60fps for HD on a modern GPU. Our technique is based on background matting, where an additional frame of the background is captured and used in recovering the alpha matte and the foreground layer. The main challenge is to compute a high-quality alpha matte, preserving strand-level hair details, while processing high-resolution images in real-time. To achieve this goal, we employ two neural networks; a base network computes a low-resolution result which is refined by a second network operating at high-resolution on selective patches. We introduce two large-scale video and image matting datasets: VideoMatte240K and PhotoMatte13K/85. Our approach yields higher quality results compared to the previous state-of-the-art in background matting, while simultaneously yielding a dramatic boost in both speed and resolution.

We present a framework for automatically reconfiguring images of street scenes by populating, depopulating, or repopulating them with objects such as pedestrians or vehicles. Applications of this method include anonymizing images to enhance privacy, generating data augmentations for perception tasks like autonomous driving, and composing scenes to achieve a certain ambiance, such as empty streets in the early morning. At a technical level, our work has three primary contributions: (1) a method for clearing images of objects, (2) a method for estimating sun direction from a single image, and (3) a way to compose objects in scenes that respects scene geometry and illumination. Each component is learned from data with minimal ground truth annotations, by making creative use of large-numbers of short image bursts of street scenes. We demonstrate convincing results on a range of street scenes and illustrate potential applications.

Removing objects from images is a challenging technical problem that is important for many applications, including mixed reality. For believable results, the shadows that the object casts should also be removed. Current inpainting-based methods only remove the object itself, leaving shadows behind, or at best require specifying shadow regions to inpaint. We introduce a deep learning pipeline for removing a shadow along with its caster. We leverage rough scene models in order to remove a wide variety of shadows (hard or soft, dark or subtle, large or thin) from surfaces with a wide variety of textures. We train our pipeline on synthetically rendered data, and show qualitative and quantitative results on both synthetic and real scenes.

Single image 3D photography enables viewers to view a still image from novel viewpoints. Recent approaches combine monocular depth networks with inpainting networks to achieve compelling results. A drawback of these techniques is the use of hard depth layering, making them unable to model intricate appearance details such as thin hair-like structures. We present SLIDE, a modular and unified system for single image 3D photography that uses a simple yet effective soft layering strategy to better preserve appearance details in novel views. In addition, we propose a novel depth-aware training strategy for our inpainting module, better suited for the 3D photography task. The resulting SLIDE approach is modular, enabling the use of other components such as segmentation and matting for improved layering. At the same time, SLIDE uses an efficient layered depth formulation that only requires a single forward pass through the component networks to produce high quality 3D photos. Extensive experimental analysis on three view-synthesis datasets, in combination with user studies on in-the-wild image collections, demonstrate superior performance of our technique in comparison to existing strong baselines while being conceptually much simpler. Project page: https://varunjampani.github.io/slide

Every time you sit in front of a TV or monitor, your face is actively illuminated by time-varying patterns of light. This paper proposes to use this time-varying illumination for synthetic relighting of your face with any new illumination condition. In doing so, we take inspiration from the light stage work of Debevec et al. [4], who first demonstrated the ability to relight people captured in a controlled lighting environment. Whereas existing light stages require expensive, room-scale spherical capture gantries and exist in only a few labs in the world, we demonstrate how to acquire useful data from a normal TV or desktop monitor. Instead of subjecting the user to uncomfortable rapidly flashing light patterns, we operate on images of the user watching a YouTube video or other standard content. We train a deep network on images plus monitor patterns of a given user and learn to predict images of that user under any target illumination (monitor pattern). Experimental evaluation shows that our method produces realistic relighting results. Video results are available at grail.cs.washington.edu/projects/Light_Stage_on_Every_Desk/.

Abstract We propose a novel method to generate fabrication blueprints from images of carpentered items. While 3D reconstruction from images is a well‐studied problem, typical approaches produce representations that are ill‐suited for computer‐aided design and fabrication applications. Our key insight is that fabrication processes define and constrain the design space for carpentered objects, and can be leveraged to develop novel reconstruction methods. Our method makes use of domain‐specific constraints to recover not just valid geometry, but a semantically valid assembly of parts, using a combination of image‐based and geometric optimization techniques. We demonstrate our method on a variety of wooden objects and furniture, and show that we can automatically obtain designs that are both easy to edit and accurate recreations of the ground truth. We further illustrate how our method can be used to fabricate a physical replica of the captured object as well as a customized version, which can be produced by directly editing the reconstructed model in CAD software.

We propose a method for creating a matte - the per-pixel foreground color and alpha - of a person by taking photos or videos in an everyday setting with a handheld camera. Most existing matting methods require a green screen background or a manually created trimap to produce a good matte. Automatic, trimap-free methods are appearing, but are not of comparable quality. In our trimap free approach, we ask the user to take an additional photo of the background without the subject at the time of capture. This step requires a small amount of foresight but is far less timeconsuming than creating a trimap. We train a deep network with an adversarial loss to predict the matte. We first train a matting network with a supervised loss on ground truth data with synthetic composites. To bridge the domain gap to real imagery with no labeling, we train another matting network guided by the first network and by a discriminator that judges the quality of composites. We demonstrate results on a wide variety of photos and videos and show significant improvement over the state of the art.

We present a new framework for sketch-based modeling and animation of 3D organic shapes that can work entirely in an intuitive 2D domain, enabling a playful, casual experience. Unlike previous sketch-based tools, our approach does not require a tedious part-based multi-view workflow with the explicit specification of an animation rig. Instead, we combine 3D inflation with a novel rigidity-preserving, layered deformation model, ARAP-L, to produce a smooth 3D mesh that is immediately ready for animation. Moreover, the resulting model can be animated from a single viewpoint --- and without the need to handle unwanted inter-penetrations, as required by previous approaches. We demonstrate the benefit of our approach on a variety of examples produced by inexperienced users as well as professional animators. For less experienced users, our single-view approach offers a simpler modeling and animating experience than working in a 3D environment, while for professionals, it offers a quick and casual workspace for ideation.

We present a method and application for animating a human subject from a single photo. E.g., the character can walk out, run, sit, or jump in 3D. The key contributions of this paper are: 1) an application of viewing and animating humans in single photos in 3D, 2) a novel 2D warping method to deform a posable template body model to fit the person's complex silhouette to create an animatable mesh, and 3) a method for handling partial self occlusions. We compare to state-of-the-art related methods and evaluate results with human studies. Further, we present an interactive interface that allows re-posing the person in 3D, and an augmented reality setup where the animated 3D person can emerge from the photo into the real world. We demonstrate the method on photos, posters, and art. The project page is at https://grail.cs.washington.edu/projects/wakeup/.

We present a system that transforms a monocular video of a soccer game into a moving 3D reconstruction, in which the players and field can be rendered interactively with a 3D viewer or through an Augmented Reality device. At the heart of our paper is an approach to estimate the depth map of each player, using a CNN that is trained on 3D player data extracted from soccer video games. We compare with state of the art body pose and depth estimation techniques, and show results on both synthetic ground truth benchmarks, and real YouTube soccer footage.

We introduce the light localization problem. A scene is illuminated by a set of unobserved isotropic point lights. Given the geometry, materials, and illuminated, appearance of the scene, the light localization problem is to completely recover the number, positions, and intensities of the lights. We first present a scene transform that identifies likely light positions. Based on this transform, we develop an iterative algorithm to locate remaining lights and determine all light intensities. We demonstrate the success of this method in a large set of 2D synthetic scenes, and show that it extends to 3D, in both synthetic scenes and real-world scenes.

Visualizing changes to indoor scenes is important for many applications. When looking for a new place to live, we want to see how the interior looks not with the current inhabitant's belongings, but with our own furniture. Before purchasing a new sofa, we want to visualize how it would look in our living room. In this paper, we present a system that takes an RGBD scan of an indoor scene and produces a scene model of the empty room, including light emitters, materials, and the geometry of the non-cluttered room. Our system enables realistic rendering not only of the empty room under the original lighting conditions, but also with various scene edits, including adding furniture, changing the material properties of the walls, and relighting. These types of scene edits enable many mixed reality applications in areas such as real estate, furniture retail, and interior design. Our system contains two novel technical contributions: a 3D radiometric calibration process that recovers the appearance of the scene in high dynamic range, and a global-illumination-aware inverse rendering framework that simultaneously recovers reflectance properties of scene surfaces and lighting properties for several light source types, including generalized point and line lights.

In this paper, we use matting to separate foreground layers from light fields captured with a plenoptic camera. We represent the input 4D light field as a 4D background light field, plus a 2D spatially varying foreground color layer with alpha. Our method can be used to both pull a foreground matte and estimate an occluded background light field. Our method assumes that the foreground layer is thin and fronto-parallel, and is composed of a limited set of colors that are distinct from the background layer colors. Our method works well for thin, translucent, and blurred foreground occluders. Our representation can be used to render the light field from novel views, handling disocclusions while avoiding common artifacts.

We address the problem of geo-registering ground-based multi-view stereo models by ground-to-aerial image matching. The main contribution is a fully automated geo-registration pipeline with a novel viewpoint-dependent matching method that handles ground to aerial viewpoint variation. We conduct large-scale experiments which consist of many popular outdoor landmarks in Rome. The proposed approach demonstrates a high success rate for the task, and dramatically outperforms state-of-the-art techniques, yielding geo-registration at pixel-level accuracy.

We present MotionMontage, a system for recording multiple motion takes of a rigid virtual object and compositing them together into a montage. Our system incorporates a Kinect-based performance capture setup that allows animators to create 3D animations by tracking the motion of a rigid physical object and mapping it in realtime onto a virtual object. The animator then temporally annotates the best parts of each take. MotionMontage merges the annotated motions into a single composite montage using a combination of dynamic time warping and optimization of a Semi-Markov Conditional Random Field. Our system also supports the creation of layered animations in which multiple objects are moving at the same time. To aid the animator in coordinating the motions of the objects we provide spatial markers which indicate the positions of previously recorded objects at user-specified points in time. We perform a user study to evaluate the perceived quality of the montages created with our system and find that viewers (including both the original animators and new viewers) generally prefer the animation montage to any individual take.

This paper leverages occluding contours (aka "internal silhouettes") to improve the performance of multi-view stereo methods. The contributions are 1) a new technique to identify free-space regions arising from occluding contours, and 2) a new approach for incorporating the resulting free-space constraints into Poisson surface reconstruction. The proposed approach outperforms state of the art MVS techniques for challenging Internet datasets, yielding dramatic quality improvements both around object contours and in surface detail.

In this paper, we propose a simple, novel plane sweep technique for refocusing plenoptic images. Rays are projected directly from the raw plenoptic image captured on the sensor into the output image plane, without computing intermediate representations such as subaperture views or epipolar images. Interpolation is performed in the output image plane using splatting. The splat kernel for each ray is adjusted adaptively, based on the refocus depth and an estimate of the depth at which that ray intersects the scene. This adaptive interpolation method antialiases out-of-focus regions, while keeping in-focus regions sharp. We test the proposed method on images from a Lytro camera and compare our results with those from the Lytro SDK. Additionally, we provide a thorough discussion of our calibration and preprocessing pipeline for this camera.

We present the first large scale system for capturing and rendering relight able scene reconstructions from massive unstructured photo collections taken under different illumination conditions and viewpoints. We combine photos taken from many sources, Flickr-Based ground-level imagery, oblique aerial views, and street view, to recover models that are significantly more complete and detailed than previously demonstrated. We demonstrate the ability to match both the viewpoint and illumination of arbitrary input photos, enabling a Visual Turing Test in which photo and rendering are viewed side-by-side and the observer has to guess which is which. While we cannot yet fool human perception, the gap is closing.

Imagining what a proposed home remodel might look like without actually performing it is challenging. We present an image-based remodeling methodology that allows real-time photorealistic visualization during both the modeling and remodeling process of a home interior. Large-scale edits, like removing a wall or enlarging a window, are performed easily and in real time, with realistic results. Our interface supports the creation of concise, parameterized, and constrained geometry, as well as remodeling directly from within the photographs. Real-time texturing of modified geometry is made possible by precomputing view-dependent textures for all faces that are potentially visible to each original camera viewpoint, blending multiple viewpoints and hole-filling when necessary. The resulting textures are stored and accessed efficiently enabling intuitive real-time realistic visualization, modeling, and editing of the building interior.

We propose a novel framework for reconstructing homogenous, transparent, refractive height-fields from a single viewpoint. The height-field is imaged against a known planar background, or sequence of backgrounds. Unlike existing approaches that do a point-by-point reconstruction - which is known to have intractable ambiguities - our method estimates and optimizes for the entire height-field at the same time. The formulation supports shape recovery from measured distortions (deflections) or directly from the images themselves, including from a single image. We report results for a variety of refractive height-fields showing significant improvement over prior art.

This paper introduces a schematic representation for architectural scenes together with robust algorithms for reconstruction from sparse 3D point cloud data. The schematic models architecture as a network of transport curves, approximating a floorplan, with associated profile curves, together comprising an interconnected set of swept surfaces. The representation is extremely concise, composed of a handful of planar curves, and easily interpretable by humans. The approach also provides a principled mechanism for interpolating a dense surface, and enables filling in holes in the data, by means of a pipeline that employs a global optimization over all parameters. By incorporating a displacement map on top of the schematic surface, it is possible to recover fine details. Experiments show the ability to reconstruct extremely clean and simple models from sparse structure-from-motion point clouds of complex architectural scenes.

We demonstrate a realtime system which infers and tracks the assembly process of a snap-together block model using a Kinect® sensor. The inference enables us to build a virtual replica of the model at every step. Tracking enables us to provide context specific visual feedback on a screen by augmenting the rendered virtual model aligned with the physical model. The system allows users to author a new model and uses the inferred assembly process to guide its recreation by others. We propose a novel way of assembly guidance where the next block to be added is rendered in blinking mode with the tracked virtual model on screen. The system is also able to detect any mistakes made and helps correct them by providing appropriate feedback. We focus on assemblies of Duplo® blocks.

We present a system for producing 3D animations using physical objects (i.e., puppets) as input. Puppeteers can load 3D models of familiar rigid objects, including toys, into our system and use them as puppets for an animation. During a performance, the puppeteer physically manipulates these puppets in front of a Kinect depth sensor. Our system uses a combination of image-feature matching and 3D shape matching to identify and track the physical puppets. It then renders the corresponding 3D models into a virtual set. Our system operates in real time so that the puppeteer can immediately see the resulting animation and make adjustments on the fly. It also provides 6D virtual camera \\rev{and lighting} controls, which the puppeteer can adjust before, during, or after a performance. Finally our system supports layered animations to help puppeteers produce animations in which several characters move at the same time. We demonstrate the accessibility of our system with a variety of animations created by puppeteers with no prior animation experience.

We present the design and implementation of new inexact Newton type Bundle Adjustment algorithms that exploit hardware parallelism for efficiently solving large scale 3D scene reconstruction problems. We explore the use of multicore CPU as well as multicore GPUs for this purpose. We show that overcoming the severe memory and bandwidth limitations of current generation GPUs not only leads to more space efficient algorithms, but also to surprising savings in runtime. Our CPU based system is up to ten times and our GPU based system is up to thirty times faster than the current state of the art methods, while maintaining comparable convergence behavior. The code and additional results are available at http://grail.cs.washington.edu/projects/mcba.

Recognizing and manipulating objects is an important task for mobile robots performing useful services in everyday environments. In this paper, we develop a system that enables a robot to grasp an object and to move it in front of its depth camera so as to build a 3D surface model of the object. We derive an information gain based variant of the next best view algorithm in order to determine how the manipulator should move the object in front of the camera. By considering occlusions caused by the robot manipulator, our technique also determines when and how the robot should re-grasp the object in order to build a complete model.

In this paper, we train a computer to select still frames from video that work well as candid portraits. Because of the subjective nature of this task, we conduct a human subjects study to collect ratings of video frames across multiple videos. Then, we compute a number of features and train a model to predict the average rating of a video frame. We evaluate our model with cross-validation, and show that it is better able to select quality still frames than previous techniques, such as simply omitting frames that contain blinking or motion blur, or selecting only smiles. We also evaluate our technique qualitatively on videos that were not part of our validation set, and were taken outdoors and under different lighting conditions.

This paper introduces an approach for enabling existing multi-view stereo methods to operate on extremely large unstructured photo collections. The main idea is to decompose the collection into a set of overlapping sets of photos that can be processed in parallel, and to merge the resulting reconstructions. This overlapping clustering problem is formulated as a constrained optimization and solved iteratively. The merging algorithm, designed to be parallel and out-of-core, incorporates robust filtering steps to eliminate low-quality reconstructions and enforce global visibility constraints. The approach has been tested on several large datasets downloaded from Flickr.com, including one with over ten thousand images, yielding a 3D reconstruction with nearly thirty million points.

This paper describes a photometric stereo method designed for surfaces with spatially-varying BRDFs, including surfaces with both varying diffuse and specular properties. Our optimization-based method builds on the observation that most objects are composed of a small number of fundamental materials by constraining each pixel to be representable by a combination of at most two such materials. This approach recovers not only the shape but also material BRDFs and weight maps, yielding accurate rerenderings under novel lighting conditions for a wide variety of objects. We demonstrate examples of interactive editing operations made possible by our approach.

We present an optimization framework for exploring gradient-domain solutions for image and video processing. The proposed framework unifies many of the key ideas in the gradient-domain literature under a single optimization formulation. Our hope is that this generalized framework will allow the reader to quickly gain a general understanding of the field and contribute new ideas of their own. We propose a novel metric for measuring local gradient saliency that identifies salient gradients that give rise to long, coherent edges, even when the individual gradients are faint. We present a general weighting scheme for gradient constraints that improves the visual appearance of results. We also provide a solution for applying gradient-domain filters to videos and video streams in a coherent manner. Finally, we demonstrate the utility of our formulation in creating effective yet simple to implement solutions for various image-processing tasks. To exercise our formulation we have created a new saliency-based sharpen filter and a pseudo image-relighting application. We also revisit and improve upon previously defined filters such as nonphotorealistic rendering, image deblocking, and sparse data interpolation over images (e.g., colorization using optimization).

Multi-view stereo (MVS) algorithms now produce reconstructions that rival laser range scanner accuracy. However, stereo algorithms require textured surfaces, and therefore work poorly for many architectural scenes (e.g., building interiors with textureless, painted walls). This paper presents a novel MVS approach to overcome these limitations for Manhattan World scenes, i.e., scenes that consists of piece-wise planar surfaces with dominant directions. Given a set of calibrated photographs, we first reconstruct textured regions using an existing MVS algorithm, then extract dominant plane directions, generate plane hypotheses, and recover per-view depth maps using Markov random fields. We have tested our algorithm on several datasets ranging from office interiors to outdoor buildings, and demonstrate results that outperform the current state of the art for such texture-poor scenes.

This paper proposes a fully automated 3D reconstruction and visualization system for architectural scenes (interiors and exteriors). The reconstruction of indoor environments from photographs is particularly challenging due to texture-poor planar surfaces such as uniformly-painted walls. Our system first uses structure-from-motion, multi-view stereo, and a stereo algorithm specifically designed for Manhattan-world scenes (scenes consisting predominantly of piece-wise planar surfaces with dominant directions) to calibrate the cameras and to recover initial 3D geometry in the form of oriented points and depth maps. Next, the initial geometry is fused into a 3D model with a novel depth-map integration algorithm that, again, makes use of Manhattan-world assumptions and produces simplified 3D models. Finally, the system enables the exploration of reconstructed environments with an interactive, image-based 3D viewer. We demonstrate results on several challenging datasets, including a 3D reconstruction and image-based walk-through of an entire floor of a house, the first result of this kind from an automated computer vision system.

We explore the use of tracked 2D object motion to enable novel approaches to interacting with video. These include moving annotations, video navigation by direct manipulation of objects, and creating an image composite from multiple video frames. Features in the video are automatically tracked and grouped in an off-line preprocess that enables later interactive manipulation. Examples of annotations include speech and thought balloons, video graffiti, path arrows, video hyperlinks, and schematic storyboards. We also demonstrate a direct-manipulation interface for random frame access using spatial constraints, and a drag-and-drop interface for assembling still images from videos. Taken together, our tools can be employed in a variety of applications including film and video editing, visual tagging, and authoring rich media such as hyperlinked video.

This paper develops a theory of frequency domain invariants in computer vision. We derive novel identities using spherical harmonics, which are the angular frequency domain analog to common spatial domain invariants such as reflectance ratios. These invariants are derived from the spherical harmonic convolution framework for reflection from a curved surface. Our identities apply in a number of canonical cases, including single and multiple images of objects under the same and different lighting conditions. One important case we consider is two different glossy objects in two different lighting environments. For this case, we derive a novel identity, independent of the specific lighting configurations or BRDFs, that allows us to directly estimate the fourth image if the other three are available. The identity can also be used as an invariant to detecttampering in the images. While this paper is primarily theoretical, it has the potential to lay the mathematical foundations for two important practical applications. First, we can develop more general algorithms for inverse rendering problems, which can directly relight and change material properties by transferring the BRDF or lighting from another object or illumination. Second, we can check the consistency of an image, to detect tampering or image splicing.

We present a system for creating and viewing interactive exploded views of complex 3D models. In our approach, a 3D input model is organized into an explosion graph that encodes how parts explode with respect to each other. We present an automatic method for computing explosion graphs that takes into account part hierarchies in the input models and handles common classes of interlocking parts. Our system also includes an interface that allows users to interactively explore our exploded views using both direct controls and higher-level interaction modes.

We present a multi-view stereo algorithm that addresses the extreme changes in lighting, scale, clutter, and other effects in large online community photo collections. Our idea is to intelligently choose images to match, both at a per-view and per-pixel level. We show that such adaptive view selection enables robust performance even with dramatic appearance variability. The stereo matching technique takes as input sparse 3D points reconstructed from structure-from-motion methods and iteratively grows surfaces from these points. Optimizing for surface normals within a photoconsistency measure significantly improves the matching results. While the focus of our approach is to estimate high-quality depth maps, we also show examples of merging the resulting depth maps into compelling scene reconstructions. We demonstrate our algorithm on standard multi-view stereo datasets and on casually acquired photo collections of famous scenes gathered from the Internet.

We present a system for authoring and viewing interactive cutaway illustrations of complex 3D models using conventions of traditional scientific and technical illustration. Our approach is based on the two key ideas that 1) cuts should respect the geometry of the parts being cut, and 2) cutaway illustrations should support interactive exploration. In our approach, an author instruments a 3D model with auxiliary parameters, which we call "rigging," that define how cutaways of that structure are formed. We provide an authoring interface that automates most of the rigging process. We also provide a viewing interface that allows viewers to explore rigged models using high-level interactions. In particular, the viewer can just select a set of target structures, and the system will automatically generate a cutaway illustration that exposes those parts. We have tested our system on a variety of CAD and anatomical models, and our results demonstrate that our approach can be used to create and view effective interactive cutaway illustrations for a variety of complex objects with little user effort.

Why do we take pictures and videos? Often, the answer is that we hope to capture moments in time, so that we can later recall and savor them once again. Digital cameras and camcorders are making it ever easier to record these moments, but often, something is lost. Photographs freeze time and space, losing the sense of motion in a scene and losing the freedom of motion available to the original viewer, and video is usually of lower resolution and finite duration, and still gives up viewpoint freedom. Furthermore, the task of sorting through the reams of image and video data that an individual records is becoming simply burdensome. In this talk, I will describe research aimed at helping the user to better capture and re-experience the moment. One approach is to build complex hardware to acquire an immersive representation of the scene, allowing virtual flythroughs and the like. The work I will describe is far less heavyhanded – it is based on handfuls of images and simple video captures and

We present a technique for computing a dense pixel correspondence between two images of a scene containing multiple large, rigid motions. We model each motion with either a homography (for planar objects) or a fundamental matrix. The various motions in the scene are first extracted by clustering an initial sparse set of correspondences between feature points; we then perform a multi-label graph cut optimization which assigns each pixel to an independent motion and computes its disparity with respect to that motion. We demonstrate our technique on several example scenes and compare our results with previous approaches.

We present an extremely simple yet robust multi-view stereo algorithm and analyze its properties. The algorithm first computes individual depth maps using a window-based voting approach that returns only good matches. The depth maps are then merged into a single mesh using a straightforward volumetric approach. We show results for several datasets, showing accuracy comparable to the best of the current state of the art techniques and rivaling more complex algorithms.

This paper presents a quantitative comparison of several multi-view stereo reconstruction algorithms. Until now, the lack of suitable calibrated multi-view image datasets with known ground truth (3D shape models) has prevented such direct comparisons. In this paper, we first survey multi-view stereo algorithms and compare them qualitatively using a taxonomy that differentiates their key properties. We then describe our process for acquiring and calibrating multiview image datasets with high-accuracy ground truth and introduce our evaluation methodology. Finally, we present the results of our quantitative comparison of state-of-the-art multi-view stereo reconstruction algorithms on six benchmark datasets. The datasets, evaluation details, and instructions for submitting new models are available online at http://vision.middlebury.edu/mview.

We present a method for visualizing short video clips in a single static image, using the visual language of storyboards. These schematic storyboards are composed from multiple input frames and annotated using outlines, arrows, and text describing the motion in the scene. The principal advantage of this storyboard representation over standard representations of video -- generally either a static thumbnail image or a playback of the video clip in its entirety -- is that it requires only a moment to observe and comprehend but at the same time retains much of the detail of the source video. Our system renders a schematic storyboard layout based on a small amount of user interaction. We also demonstrate an interaction technique to scrub through time using the natural spatial dimensions of the storyboard. Potential applications include video editing, surveillance summarization, assembly instructions, composition of graphic novels, and illustration of camera technique for film studies.

This paper describes a photometric stereo method designed for surfaces with spatially-varying BRDFs, including surfaces with both varying diffuse and specular properties. Our method builds on the observation that most objects are composed of a small number of fundamental materials. This approach recovers not only the shape but also material BRDFs and weight maps, yielding compelling results for a wide variety of objects. We also show examples of interactive lighting and editing operations made possible by our method.

This paper describes a mostly automatic method for taking the output of a single panning video camera and creating a panoramic video texture (PVT): a video that has been stitched into a single, wide field of view and that appears to play continuously and indefinitely. The key problem in creating a PVT is that although only a portion of the scene has been imaged at any given time, the output must simultaneously portray motion throughout the scene. Like previous work in video textures, our method employs min-cut optimization to select fragments of video that can be stitched together both spatially and temporally. However, it differs from earlier work in that the optimization must take place over a much larger set of data. Thus, to create PVTs, we introduce a dynamic programming step, followed by a novel hierarchical min-cut optimization algorithm. We also use gradient-domain compositing to further smooth boundaries between video fragments. We demonstrate our results with an interactive viewer in which users can interactively pan and zoom on high-resolution PVTs.

In this paper, we explore the problem of enhancing still pictures with subtly animated motions. We limit our domain to scenes containing passive elements that respond to natural forces in some fashion. We use a semi-automatic approach, in which a human user segments the scene into a series of layers to be individually animated. Then, a "stochastic motion texture" is automatically synthesized using a spectral method, i.e., the inverse Fourier transform of a filtered noise spectrum. The motion texture is a time-varying 2D displacement map, which is applied to each layer. The resulting warped layers are then recomposited to form the animated frames. The result is a looping video texture created from a single still image, which has the advantages of being more controllable and of generally higher image quality and resolution than a video texture created from a video source. We demonstrate the technique on a variety of photographs and paintings.

We describe an interactive, computer-assisted framework for combining parts of a set of photographs into a single composite picture, a process we call "digital photomontage." Our framework makes use of two techniques primarily: graph-cut optimization, to choose good seams within the constituent images so that they can be combined as seamlessly as possible; and gradient-domain fusion, a process based on Poisson equations, to further reduce any remaining visible artifacts in the composite. Also central to the framework is a suite of interactive tools that allow the user to specify a variety of high-level image objectives, either globally across the image, or locally through a painting-style interface. Image objectives are applied independently at each pixel location and generally involve a function of the pixel values (such as "maximum contrast") drawn from that same location in the set of source images. Typically, a user applies a series of image objectives iteratively in order to create a finished composite. The power of this framework lies in its generality; we show how it can be used for a wide variety of applications, including "selective composites" (for instance, group photos in which everyone looks their best), relighting, extended depth of field, panoramic stitching, clean-plate production, stroboscopic visualization of movement, and time-lapse mosaics.

We present an end-to-end system that goes from video sequences to high resolution, editable, dynamically controllable face models. The capture system employs synchronized video cameras and structured light projectors to record videos of a moving face from multiple viewpoints. A novel spacetime stereo algorithm is introduced to compute depth maps accurately and overcome over-fitting deficiencies in prior work. A new template fitting and tracking procedure fills in missing data and yields point correspondence across the entire sequence without using markers. We demonstrate a data-driven, interactive method for inverse kinematics that draws on the large set of fitted templates and allows for posing new expressions by dragging surface points directly. Finally, we describe new tools that model the dynamics in the input sequence to enable new animations, created via key-framing or texture-synthesis techniques.

This paper extends the traditional binocular stereo problem into the spacetime domain, in which a pair of video streams is matched simultaneously instead of matching pairs of images frame by frame. Almost any existing stereo algorithm may be extended in this manner simply by replacing the image matching term with a spacetime term. By utilizing both spatial and temporal appearance variation, this modification reduces ambiguity and increases accuracy. Three major applications for spacetime stereo are proposed in this paper. First, spacetime stereo serves as a general framework for structured light scanning and generates high quality depth maps for static scenes. Second, spacetime stereo is effective for a class of natural scenes, such as waving trees and flowing water, which have repetitive textures and chaotic behaviors and are challenging for existing stereo algorithms. Third, the approach is one of very few existing methods that can robustly reconstruct objects that are moving and deforming over time, achieved by use of oriented spacetime windows in the matching procedure. Promising experimental results in the above three scenarios are demonstrated.

We present an algorithm for computing optical flow, shape, motion, lighting, and albedo from an image sequence of a rigidly-moving Lambertian object under distant illumination. The problem is formulated in a manner that subsumes structure from motion, multiview stereo, and photometric stereo as special cases. The algorithm utilizes both spatial and temporal intensity variation as cues: the former constrains flow and the latter constrains surface orientation; combining both cues enables dense reconstruction of both textured and textureless surfaces. The algorithm works by iteratively estimating affine camera parameters, illumination, shape, and albedo in an alternating fashion. Results are demonstrated on videos of hand-held objects moving in front of a fixed light and camera.

We develop a novel method for fitting high-resolution template meshes to detailed human body range scans with sparse 3D markers. We formulate an optimization problem in which the degrees of freedom are an affine transformation at each template vertex. The objective function is a weighted combination of three measures: proximity of transformed vertices to the range data, similarity between neighboring transformations, and proximity of sparse markers at corresponding locations on the template and target surface. We solve for the transformations with a non-linear optimizer, run at two resolutions to speed convergence. We demonstrate reconstruction and consistent parameterization of 250 human body models. With this parameterized set, we explore a variety of applications for human body modeling, including: morphing, texture transfer, statistical analysis of shape, model fitting from sparse markers, feature analysis to modify multiple correlated parameters (such as the weight and height of an individual), and transfer of surface detail and animation controls from a template to fitted models.

In this paper, we describe a method for extracting shadows from one natural scene and inserting them into another. We develop physically-based shadow matting and compositing equations and use these to pull a shadow matte from a source scene in which the shadow is cast onto an arbitrary planar background. We then acquire the photometric and geometric properties of the target scene by sweeping oriented linear shadows (cast by a straight object) across it. From these shadow scans, we can construct a shadow displacement map without requiring camera or light source calibration. This map can then be used to deform the original shadow matte. We demonstrate our approach for both indoor scenes with controlled lighting and for outdoor scenes using natural lighting.

This paper presents an example-based method for calculating skeleton-driven body deformations. Our example data consists of range scans of a human body in a variety of poses. Using markers captured during range scanning, we construct a kinematic skeleton and identify the pose of each scan. We then construct a mutually consistent parameterization of all the scans using a posable subdivision surface template. The detail deformations are represented as displacements from this surface, and holes are filled smoothly within the displacement maps. Finally, we combine the range scans using k -nearest neighbor interpolation in pose space. We demonstrate results for a human upper body with controllable pose, kinematics, and underlying surface shape.

This paper presents a framework for the skeleton-driven animation of elastically deformable characters. A character is embedded in a coarse volumetric control lattice, which provides the structure needed to apply the finite element method. To incorporate skeletal controls, we introduce line constraints along the bones of simple skeletons. The bones are made to coincide with edges of the control lattice, which enables us to apply the constraints efficiently using algebraic methods. To accelerate computation, we associate regions of the volumetric mesh with particular bones and perform locally linearized simulations, which are blended at each time step. We define a hierarchical basis on the control lattice, so for detailed interactions the simulation can adapt the level of detail. We demonstrate the ability to animate complex models using simple skeletons and coarse volumetric meshes in a manner that simulates secondary motions at interactive rates.

This paper describes a new framework for video matting, the process of pulling a high-quality alpha matte and foreground from a video sequence. The framework builds upon techniques in natural image matting, optical flow computation, and background estimation. User interaction is comprised of garbage matte specification if background estimation is needed, and hand-drawn keyframe segmentations into "foreground," "background" and "unknown". The segmentations, called trimaps, are interpolated across the video volume using forward and backward optical flow. Competing flow estimates are combined based on information about where flow is likely to be accurate. A Bayesian matting technique uses the flowed trimaps to yield high-quality mattes of moving foreground elements with complex boundaries filmed by a moving camera. A novel technique for smoke matte extraction is also demonstrated.

This paper proposes a new Bayesian framework for solving the matting problem, i.e. extracting a foreground element from a background image by estimating an opacity for each pixel of the foreground element. Our approach models both the foreground and background color distributions with spatially-varying sets of Gaussians, and assumes a fractional blending of the foreground and background colors to produce the final output. It then uses a maximum-likelihood criterion to estimate the optimal opacity, foreground and background simultaneously. In addition to providing a principled approach to the matting problem, our algorithm effectively handles objects with intricate boundaries, such as hair strands and fur, and provides an improvement over existing techniques for these difficult cases.

This paper describes a new framework for processing images by example, called “image analogies.” The framework involves two stages: a design phase, in which a pair of images, with one image purported to be a “filtered” version of the other, is presented as “training data”; and an application phase, in which the learned filter is applied to some new target image in order to create an “analogous” filtered result. Image analogies are based on a simple multi-scale autoregression, inspired primarily by recent results in texture synthesis. By choosing different types of source image pairs as input, the framework supports a wide variety of “image filter” effects, including traditional image filters, such as blurring or embossing; improved texture synthesis, in which some textures are synthesized with higher quality than by previous approaches; super-resolution, in which a higher-resolution image is inferred from a low-resolution source; texture transfer, in which images are “texturized” with some arbitrary source texture; artistic filters, in which various drawing and painting styles are synthesized based on scanned real-world examples; and texture-by-numbers, in which realistic scenes, composed of a variety of textures, are created using a simple painting interface.

Environment matting is a generalization of traditional bluescreen matting. By photographing an object in front of a sequence of structured light backdrops, a set of approximate light-transport paths through the object can be computed. The original environment matting research chose a middle ground—using a moderate number of photographs to produce results that were reasonably accurate for many objects. In this work, we extend the technique in two opposite directions: recovering a more accurate model at the expense of using additional structured light backdrops, and obtaining a simplified matte using just a single backdrop. The first extension allows for the capture of complex and subtle interactions of light with objects, while the second allows for video capture of colorless objects in motion.

A surface light field is a function that assigns a color to each ray originating on a surface. Surface light fields are well suited to constructing virtual images of shiny objects under complex lighting conditions. This paper presents a framework for construction, compression, interactive rendering, and rudimentary editing of surface light fields of real objects. Generalization of vector quantization and principal component analysis are used to construct a compressed representation of an object's surface light field from photographs and range scans. A new rendering algorithm achieves interactive rendering of images from the compressed representation, incorporating view-dependent geometric level-of-detail control. The surface light field representation can also be directly edited to yield plausible surface light fields for small changes in surface geometry and reflectance properties.

This paper introduces a new process, environment matting,which captures not just a foreground object and its traditional opacity matte from a real-world scene, but also a description of how that object refracts and reflects light, which we call an environment matte. The foreground object can then be placed in a new environment, using environment compositing, where it will refract and reflect light from that scene. Objects captured in this way exhibit not only specular but glossy and translucent effects, as well as selective attenuation and scattering of light according to wavelength. Moreover, the environment compositing process, which can be performed largely with texture mapping operations, is fast enough to run at interactive speeds on a desktop PC. We compare our results to photos of the same objects in real scenes. Applications of this work include the relighting of objects for virtual and augmented reality, more realistic 3D clip art, and interactive lighting design. CR Catego...

A number of techniques have been developed for reconstructing surfaces by integrating groups of aligned range images. A desirable set of properties for such algorithms includes: incremental updating, representation of directional uncertainty, the ability to fill gaps in the reconstruction, and robustness in the presence of outliers. Prior algorithms possess subsets of these properties. In this paper, we present a volumetric method for integrating range images that possesses all of these properties. Our volumetric representation consists of a cumulative weighted signed distance function. Working with one range image at a time, we first scan-convert it to a distance function, then combine this with the data already acquired using a simple additive scheme. To achieve space efficiency, we employ a run-length encoding of the volume. To achieve time efficiency, we resample the range image to align with the voxel grid and traverse the range and voxel scanlines synchronously. We generate the final manifold by extracting an isosurface from the volumetric grid. We show that under certain assumptions, this isosurface is optimal in the least squares sense. To fill gaps in the model, we tessellate over the boundaries between regions seen to be empty and regions never observed. Using this method, we are able to integrate a large number of range images (as many as 70) yielding seamless, high-detail models of up to 2.6 million triangles. CR Categories: I.3.5 [Computer Graphics] Computational Geometry and Object Modeling Additional keywords: Surface fitting, three-dimensional shape recovery, range image integration, isosurface extraction 1

The standard methods for extracting range data from optical triangulation scanners are accurate only for planar objects of uniform reflectance illuminated by an incoherent source. Using these methods, curved surfaces, discontinuous surfaces, and surfaces of varying reflectance cause systematic distortions of the range data. Coherent light sources such as lasers introduce speckle artifacts that further degrade the data. We present a new ranging method based on analyzing the time evolution of the structured light reflections. Using our spacetime analysis, we can correct for each of these artifacts, thereby attaining significantly higher accuracy using existing technology. We present results that demonstrate the validity of our method using a commercial laser stripe triangulation scanner.>

A high-performance parallel signal processing system for use in a real-time environment is described. The system includes analog input and output conditioners and data converters, digital signal processing (DSP) elements, and a system control computer. Expanding requirements and advances in DSP technology have resulted in the evolution of several generations of the system, each more powerful than the previous one. The hallmark of the newest system is an ensemble of DSP boards, each containing eight AT&T DSP32C processors. Custom architectures and algorithms were developed to allow the DSP chips to execute filters with more taps per second than processor cycles.>