Heung-Yeung Shum

Stereo matching using belief propagation

In this paper, we formulate the stereo matching problem as a Markov network and solve it using Bayesian belief propagation. The stereo Markov network consists of three coupled Markov random fields that model the following: a smooth field for depth/disparity, a line process for depth discontinuity, and a binary process for occlusion. After eliminating the line process and the binary process by introducing two robust functions, we apply the belief propagation algorithm to obtain the maximum a posteriori (MAP) estimation in the Markov network. Other low-level visual cues (e.g., image segmentation) can also be easily incorporated in our stereo model to obtain better stereo results. Experiments demonstrate that our methods are comparable to the state-of-the-art stereo algorithms for many test cases.

Learning to Detect a Salient Object

In this paper, we study the salient object detection problem for images. We formulate this problem as a binary labeling task where we separate the salient object from the background. We propose a set of novel features, including multiscale contrast, center-surround histogram, and color spatial distribution, to describe a salient object locally, regionally, and globally. A conditional random field is learned to effectively combine these features for salient object detection. Further, we extend the proposed approach to detect a salient object from sequential images by introducing the dynamic salient features. We collected a large image database containing tens of thousands of carefully labeled images by multiple users and a video segment database, and conducted a set of experiments over them to demonstrate the effectiveness of the proposed approach.

Lazy snapping

In this paper, we present Lazy Snapping, an interactive image cutout tool. Lazy Snapping separates coarse and fine scale processing, making object specification and detailed adjustment easy. Moreover, Lazy Snapping provides instant visual feedback, snapping the cutout contour to the true object boundary efficiently despite the presence of ambiguous or low contrast edges. Instant feedback is made possible by a novel image segmentation algorithm which combines graph cut with pre-computed over-segmentation. A set of intuitive user interface (UI) tools is designed and implemented to provide flexible control and editing for the users. Usability studies indicate that Lazy Snapping provides a better user experience and produces better segmentation results than the state-of-the-art interactive image cutout tool, Magnetic Lasso in Adobe Photoshop.

Image deblurring with blurred/noisy image pairs

Taking satisfactory photos under dim lighting conditions using a hand-held camera is challenging. If the camera is set to a long exposure time, the image is blurred due to camera shake. On the other hand, the image is dark and noisy if it is taken with a short exposure time but with a high camera gain. By combining information extracted from both blurred and noisy images, however, we show in this paper how to produce a high quality image that cannot be obtained by simply denoising the noisy image, or deblurring the blurred image alone. Our approach is image deblurring with the help of the noisy image. First, both images are used to estimate an accurate blur kernel, which otherwise is difficult to obtain from a single blurred image. Second, and again using both images, a residual deconvolution is proposed to significantly reduce ringing artifacts inherent to image deconvolution. Third, the remaining ringing artifacts in smooth image regions are further suppressed by a gain-controlled deconvolution process. We demonstrate the effectiveness of our approach using a number of indoor and outdoor images taken by off-the-shelf hand-held cameras in poor lighting environments.

Creating full view panoramic image mosaics and environment maps

This paper presents a novel approach to creating full view panoramic mosaics from image sequences. Unlike current panoramic stitching methods, which usually require pure horizontal camera panning, our system does not require any controlled motions or constraints on how the images are taken (as long as there is no strong motion parallax). For example, images taken from a hand-held digital camera can be stitched seamlessly into panoramic mosaics. Because we represent our image mosaics using a set of transforms, there are no singularity problems such as those existing at the top and bottom of cylindrical or spherical maps. Our algorithm is fast and robust because it directly recovers 3D rotations instead of general 8 parameter planar perspective transforms. Methods to recover camera focal length are also presented. We also present an algorithm for efficiently extracting environment maps from our image mosaics. By mapping the mosaic onto an artibrary texture-mapped polyhedron surrounding the origin, we can explore the virtual environment using standard 3D graphics viewers and hardware without requiring special-purpose players. CR

Plenoptic sampling

This paper studies the problem of plenoptic sampling in image-based rendering (IBR). From a spectral analysis of light field signals and using the sampling theorem, we mathematically derive the analytical functions to determine the minimum sampling rate for light field rendering. The spectral support of a light field signal is bounded by the minimum and maximum depths only, no matter how complicated the spectral support might be because of depth variations in the scene. The minimum sampling rate for light field rendering is obtained by compacting the replicas of the spectral support of the sampled light field within the smallest interval. Given the minimum and maximum depths, a reconstruction filter with an optimal and constant depth can be designed to achieve anti-aliased light field rendering. Plenoptic sampling goes beyond the minimum number of images needed for anti-aliased light field rendering. More significantly, it utilizes the scene depth information to determine the minimum sampling curve in the joint image and geometry space. The minimum sampling curve quantitatively describes the relationship among three key elements in IBR systems: scene complexity (geometrical and textural information), the number of image samples, and the output resolution. Therefore, plenoptic sampling bridges the gap between image-based rendering and traditional geometry-based rendering. Experimental results demonstrate the effectiveness of our approach.

Real-time texture synthesis by patch-based sampling

We present an algorithm for synthesizing textures from an input sample. This patch-based sampling algorithm is fast and it makes high-quality texture synthesis a real-time process. For generating textures of the same size and comparable quality, patch-based sampling is orders of magnitude faster than existing algorithms. The patch-based sampling algorithm works well for a wide variety of textures ranging from regular to stochastic. By sampling patches according to a nonparametric estimation of the local conditional MRF density function, we avoid mismatching features across patch boundaries. We also experimented with documented cases for which pixel-based nonparametric sampling algorithms cease to be effective but our algorithm continues to work well.

Rendering with concentric mosaics

An image based system and process for rendering novel views of a real or synthesized 3D scene based on a series of concentric mosaics depicting the scene. In one embodiment, each concentric mosaic represents a collection of consecutive slit images of the surrounding 3D scene taken from a different viewpoint tangent to a circle on a plane within the scene. Novel views from viewpoints within circular regions of the aforementioned circle plane defined by the concentric mosaics are rendered using these concentric mosaics. Specifically, a slit image can be identified by a ray originating at its viewpoint on the circle plane and extending toward the longitudinal midline of the slit image. Each of the rays associated with the slit images needed to construct a novel view will either coincide with one of the rays associated with a previously captured slit image, or it will pass between two of the concentric circles on the circle plane. If it coincides, then the previously captured slit image associated with the coinciding ray can be used directly to construct part of the novel view. If the ray passes between two of the concentric circles of the plane, then the needed slit image is interpolated using the two previously captured slit images associated with the rays originating from the adjacent concentric circles that are parallel to the non-coinciding ray. If the objects in the 3D scene are close to the camera, depth correction is applied to reduce image distortion for pixels located above and below the circle plane. In another embodiment, a single camera is used to capture a sequence of images. Each image includes image data that has a ray direction associated therewith. To render an image at a novel viewpoint, multiple ray directions from the novel viewpoint are chosen. Image data is combined from the sequence of images by selecting image data that has a ray direction substantially aligning with the ray direction from the novel viewpoint.

Poisson matting

In this paper, we formulate the problem of natural image matting as one of solving Poisson equations with the matte gradient field. Our approach, which we call Poisson matting, has the following advantages. First, the matte is directly reconstructed from a continuous matte gradient field by solving Poisson equations using boundary information from a user-supplied trimap. Second, by interactively manipulating the matte gradient field using a number of filtering tools, the user can further improve Poisson matting results locally until he or she is satisfied. The modified local result is seamlessly integrated into the final result. Experiments on many complex natural images demonstrate that Poisson matting can generate good matting results that are not possible using existing matting techniques.

Image completion with structure propagation

In this paper, we introduce a novel approach to image completion, which we call structure propagation. In our system, the user manually specifies important missing structure information by extending a few curves or line segments from the known to the unknown regions. Our approach synthesizes image patches along these user-specified curves in the unknown region using patches selected around the curves in the known region. Structure propagation is formulated as a global optimization problem by enforcing structure and consistency constraints. If only a single curve is specified, structure propagation is solved using Dynamic Programming. When multiple intersecting curves are specified, we adopt the Belief Propagation algorithm to find the optimal patches. After completing structure propagation, we fill in the remaining unknown regions using patch-based texture synthesis. We show that our approach works well on a number of examples that are challenging to state-of-the-art techniques.