Abstract. In this paper, we propose a novel type of explicit image filter- guided filter. Derived from a local linear model, the guided filter generates the filtering output by considering the content of a guidance image, which can be the input image itself or another different image. The guided filter can perform as an edge-preserving smoothing operator like the popular bilateral filter [1], but has better behavior near the edges. It also has a theoretical connection with the matting Laplacian matrix [2], so is a more generic concept than a smoothing operator and can better utilize the structures in the guidance image. Moreover, the guidedfilterhasafastandnon-approximatelinear-time algorithm, whose computational complexity is independent of the filtering kernel size. We demonstrate that the guided filter is both effective and efficient in a great variety of computer vision and computer graphics applications including noise reduction, detail smoothing/enhancement, HDR compression, image matting/feathering, haze removal, and joint upsampling. 1

Many useful algorithms for processing images and geometry fall under the general framework of high-dimensional Gaussian filtering. This family of algorithms includes bilateral filtering and non-local means. We propose a new way to perform such filters using the permutohedral lattice, which tessellates high-dimensional space with uniform simplices. Our algorithm is the first implementation of a high-dimensional Gaussian filter that is both linear in input size and polynomial in dimensionality. Furthermore it is parameter-free, apart from the filter size, and achieves a consistently high accuracy relative to ground truth (> 45 dB). We use this to demonstrate a number of interactive-rate applications of filters in as high as eight dimensions.

We present a flexible method for fusing information from optical and range sensors based on an accelerated highdimensional filtering approach. Our system takes as input a sequence of monocular camera images as well as a stream of sparse range measurements as obtained from a laser or other sensor system. In contrast with existing approaches, we do not assume that the depth and color data streams have the same data rates or that the observed scene is fully static. Our method produces a dense, high-resolution depth map of the scene, automatically generating confidence values for every interpolated depth point. We describe how to integrate priors on object motion and appearance and how to achieve an efficient implementation using parallel processing hardware such as GPUs. 1.

Figure 1: Affinity-based edit propagation methods such as [An and Pellacini 2008] allow one to change the appearance of an image or video (e.g., the color of the bird here) using only a few strokes, yet consuming prohibitive amount of time and memory for large data (e.g., 48 minutes and 23GB for this video containing 61M pixels). Our approximation scheme drastically reduces the cost of edit propagation methods (to 8 seconds and 22MB in this example) by exploring adaptive clustering in the affinity space. Video courtesy of BBC Motion Gallery (UK). Image/video editing by strokes has become increasingly popular due to the ease of interaction. Propagating the user inputs to the rest of the image/video, however, is often time and memory consuming especially for large data. We propose here an efficient scheme that allows affinity-based edit propagation to be computed on data containing tens of millions of pixels at interactive rate (in matter of seconds). The key in our scheme is a novel means for approximately solving the optimization problem involved in edit propagation, using adaptive clustering in a high-dimensional, affinity space. Our approximation significantly reduces the cost of existing affinitybased propagation methods while maintaining visual fidelity, and enables interactive stroke-based editing even on high resolution images and long video sequences using commodity computers. 1

Figure 1: Repeated element detection and manipulation. (Left-to-right) Original image with user scribbles to indicate an object template (red) and background (green); repeated instances detected, completed, dense correspondence established, and ordered in layers; fish in the original image replaced by a different kind of fish from a reference image (top-right inset); rearranged fishes. Repeated elements are ubiquitous and abundant in both manmade and natural scenes. Editing such images while preserving the repetitions and their relations is nontrivial due to overlap, missing parts, deformation across instances, illumination variation, etc. Manually enforcing such relations is laborious and error-prone. We propose a novel framework where user scribbles are used to guide detection and extraction of such repeated elements. Our detection process, which is based on a novel boundary band method, robustly extracts the repetitions along with their deformations. The algorithm only considers the shape of the elements, and ignores similarity based on color, texture, etc. We then use topological sorting to establish a partial depth ordering of overlapping repeated instances. Missing parts on occluded instances are completed using information from other instances. The extracted repeated instances can then be seamlessly edited and manipulated for a variety of high level tasks that are otherwise difficult to perform. We demonstrate the versatility of our framework on a large set of inputs of varying complexity, showing applications to image rearrangement, edit transfer, deformation propagation, and instance replacement. image editing, shape-aware manipulation, edit propa-Keywords: gation

Figure 1: Two camouflage images produced by our technique. The left and right images have seven and four camouflaged objects, respectively, at various levels of difficulty. By removing distinguishable elements from the camouflaged objects we make feature search difficult, forcing the viewers to use conjunction search, a serial and delayed procedure. (Please zoom in for a better effect. Answer keys are on the last page.) Camouflage images contain one or more hidden figures that remain imperceptible or unnoticed for a while. In one possible explanation, the ability to delay the perception of the hidden figures is attributed to the theory that human perception works in two main phases: feature search and conjunction search. Effective camouflage images make feature based recognition difficult, and thus force the recognition process to employ conjunction search, which takes considerable effort and time. In this paper, we present a technique for creating camouflage images. To foil the feature search, we remove the original subtle texture details of the hidden figures and replace them by that of the surrounding apparent image. To leave an appropriate degree of clues for the conjunction search, we compute and assign new tones to regions in the embedded figures by performing an optimization between two conflicting terms, which we call immersion and standout, corresponding to hiding and leaving clues, respectively. We show a large number of camouflage images generated by our technique, with or without user guidance. We have tested the quality of the images in an extensive user study, showing a good control of the difficulty levels. 1

The aim of this self-contained supplement is to rigorously establish the mathematical properties of the permutohedral lattice, to adapt it for the use in Gaussian filtering, and to further justify its use. For discussion of a CPU implementation, applications and empirical comparison to other methods, see [ABD10]. The rest of the paper is organized as follows: in Section 1, we give preliminary definitions of relevant terms and notation; in Section 2, we formally define the permutohedral lattice in several ways and prove useful structural properties stemming from the definition; in Section 3, we suggest criteria for picking the ideal lattice with which to perform Gaussian filtering, and present an argument that the permutohedral lattice is the most appropriate choice within a large family of lattices; in Section 4, we discuss how to perform each step of the Gaussian filter using the permutohedral lattice. Finally, in Section 5, we analyze the overall algorithm and compare it to the implementation on Z d in [PD09]. 1. Preliminaries We begin by declaring the notation used hereafter in Table 1. Other standard notation such as R,Z applies. d The dimensionality of the underlying space. ⃗x A vector of dimensionality d+1, unless specified otherwise: {x0,x1,...,xd}. ⃗1 The d + 1-dimensional vector {1,1,...,1}. Hd The d-dimensional hyperplane

Most state-of-the-art techniques for multi-class image segmentation and labeling use conditional random fields defined over pixels or image regions. While regionlevel models often feature dense pairwise connectivity, pixel-level models are considerably larger and have only permitted sparse graph structures. In this paper, we consider fully connected CRF models defined on the complete set of pixels in an image. The resulting graphs have billions of edges, making traditional inference algorithms impractical. Our main contribution is a highly efficient approximate inference algorithm for fully connected CRF models in which the pairwise edge potentials are defined by a linear combination of Gaussian kernels. Our experiments demonstrate that dense connectivity at the pixel level substantially improves segmentation and labeling accuracy. 1

We describe a set of image editing and viewing tools that explicitly take into account the resolution of the display on which the image is viewed. Our approach is twofold. First, we design editing tools that process only the visible data, which is useful for images larger than the display. This encompasses cases such as multi-image panoramas and highresolution medical data. Second, we propose an adaptive way to set viewing parameters such brightness and contrast. Because we deal with very large images, different locations and scales often require different viewing parameters. We let users set these parameters at a few places and interpolate satisfying values everywhere else. We demonstrate the efficiency of our approach on different display and image sizes. Since the computational complexity to render a view depends on the display resolution and not the actual input image resolution, we achieve interactive image editing even on a 16 gigapixel image. 1.

Figure 1: The figure shows a pair of 4K video frames (a,b) and the corresponding optical flow result (d). Our new SimpleFlow algorithm computes the optical flow using only local operations that can be efficiently implemented on parallel architectures such as GPUs. Further, it concentrates computation where motion actually occurs, in black in (c), and uses linearly interpolation to estimate the flow in other regions, in gray in (c). This enables the computation of accurate optical maps in a reasonable amount of time (d). We show that this strategy makes the running time grow sublinearly with the frame resolution (e). This enables the processing of high-definition videos up to the 4K movie resolution in which each frame has 9 megapixels. Optical flow is a critical component of video editing applications, e.g. for tasks such as object tracking, segmentation, and selection. In this paper, we propose an optical flow algorithm called SimpleFlow whose running times increase sublinearly in the number of pixels. Central to our approach is a probabilistic representation of the motion flow that is computed using only local evidence and without resorting to global optimization. To estimate the flow in image regions where the motion is smooth, we use a sparse set of samples only, thereby avoiding the expensive computation inherent in traditional dense algorithms. We show that our results can be used as is for a variety of video editing tasks. For applications where accuracy is paramount, we use our result to bootstrap a global optimization. This significantly reduces the running times of such methods without sacrificing accuracy. We also demonstrate that the SimpleFlow algorithm can process HD and 4K footage in reasonable times. 1.