No abstract

Computational visual attention systems detect regions of interest in images. These systems have a broad range of applications in areas such as computer vision, computational aesthetics, and nonphotorealistic rendering. However, almost all the systems to date are designed for low dynamic range (LDR) images and may not be suitable for analyzing saliency in high dynamic range (HDR) images. We propose a novel algorithm for saliency analysis of HDR images that is based on virtual photographs. Taking virtual photographs is the inverse process of generating HDR images from multiple LDR exposures, and the virtual photograph sequence has the capacity to more comprehensively reveal salient content in HDR images. We demonstrate that our method can produce more consistently reliable results than existing methods.

The edge-based level set model gives no satisfactory results for images with weak edge, and the region-based model performs poorly for intensity inhomogeneity images. In this paper, we propose an improved region-based level set model that integrates both the gradient information and the region information. The proposed model defines a novel external energy term, which consists of gradient information and signed pressure forces function. In order to eliminate the re-initialization procedure of traditional level set model, an internal energy term is also introduced for the level set function to maintain signed distance function. Compared with traditional models, our model is more robust against images with weak edge and intensity inhomogeneity. Experiments on liver segmentation from abdominal CT images demonstrate the effectiveness and accuracy of the proposed method.In recent years, the composition of liver has become a vital component in the diagnosis of liver disease and surgery planning. The liver segmentation based on computer tomography (CT) image plays a more and more important role in the field of medical image processing and computer-aided diagnosis (CAD) [1] . However, liver in the CT image is always connected to some tissues such as stomach and spleen, which makes precise liver image segmentation very challenging.Several methods are commonly adopted by researchers to segment the liver region with CT imaging, such as regional growth and snake model [2] . These methods have certain limitations. For instance, regional growth segmentation method will result in over-segmentation [3] or wrong-segmentation if the growing rule is unsuitable. Also, the sensitivity to the initial contour makes it difficult for the snake model based segmentation method to get satisfying results [4] . Since the level set model proposed by Osher et al [5,6] can effectively handle the topology change, level set model based segmentation method has become a hot topic in liver image segmentation [7] .It is well known that the edge-based [8] model and region-based [9] model are the two main types of level set models. The edge-based model builds an edge stopping function using image edge information, which can drive the contour towards the object boundaries. However, for images with intense noise or weak edge, edge stopping function based on image gradient can hardly stop at the

No abstract

No abstract