In order to improve the stability of video segmentation tasks,an algorithm is proposed which combines the graph cut algorithm with support vector machine.Firstly,the algorithm keeps the background static in adjacent frames with a preprocessing step.Then,the first frame is segmented with the mean shift algorithm.After that,a support vector machine is trained with sampled pixel features (i.e.positions and color information) from the segmentation of previous frames.It then outputs corresponding probability of pixels,which is regarded as energy of graph cut algorithms to output stable segmentations.Experiments show that the proposed algorithm performs better than traditional graph cut video segmentation algorithms from the perspective of temporal coherence.Meanwhile,the proposed algorithm could be combined with other existing video segmentation algorithms by importing the probability outputted by support vector machines.

We present a novel artistic-verisimilitude driven system for watercolor rendering of images and photos. Our system achieves realistic simulation of a set of important characteristics of watercolor paintings that have not been well implemented before. Specifically, we designed several image filters to achieve: 1) watercolor-specified color transferring; 2) saliency-based level-of-detail drawing; 3) hand tremor effect due to human neural noise; and 4) an artistically controlled wet-in-wet effect in the border regions of different wet pigments. A user study indicates that our method can produce watercolor results of artistic verisimilitude better than previous filter-based or physical-based methods. Furthermore, our algorithm is efficient and can easily be parallelized, making it suitable for interactive image watercolorization.

In this paper, we propose a system for flexible face replacement in photos that does not require human intervention. The proposed replacement algorithm is able to efficiently paste one person׳s face into another׳s with different head poses and/or hairstyles. This work can enhance the quality of group photos that contain human faces, as they are seldom flawless. Previous methods require complex user interaction or auxiliary data such as a 3D head model library or a photo library, while our system does not. Given a photo, the faces are automatically detected and transformed into similar poses by fitting corresponding facial features. We then construct an energy function and optimize it for a better face contour between the transformed source face and the target one, followed by applying seamless cloning. Our system removes hair before face transfer and adds it back afterwards, maintaining the same hairstyles. We demonstrate that our system produces high-quality results without manual tuning of parameters.

Blue noise sampling is an important component in many graphics applications, but existing techniques consider mainly the spatial positions of samples, making them less effective when handling problems with non-spatial features. Examples include biological distribution in which plant spacing is influenced by non-positional factors such as tree type and size, photon mapping in which photon flux and direction are not a direct function of the attached surface, and point cloud sampling in which the underlying surface is unknown a priori. These scenarios can benefit from blue noise sample distributions, but cannot be adequately handled by prior art. Inspired by bilateral filtering, we propose a bilateral blue noise sampling strategy. Our key idea is a general formulation to modulate the traditional sample distance measures, which are determined by sample position in spatial domain, with a similarity measure that considers arbitrary per sample attributes. This modulation leads to the notion of bilateral blue noise whose properties are influenced by not only the uniformity of the sample positions but also the similarity of the sample attributes. We describe how to incorporate our modulation into various sample analysis and synthesis methods, and demonstrate applications in object distribution, photon density estimation, and point cloud sub-sampling.