THESIS
2006
xiii, 78 leaves : ill. ; 30 cm
Abstract
Recently, medical image visualization has become a hot research topic and many novel ideas have been proposed for different practical applications. The general awareness of health and the continuously increasing number of patients suffering from various kinds of diseases show the importance of this research. Based on the observation of the severity of vascular diseases, we dedicate our effort to the research topic of vas-cular image visualization. We believe that effective visualization of images is critical to physicians for proper analysis and diagnosis of existing pathologies. Different from previous work related to medical visualization, we mainly focus on how to visualize angiograms more accurately and effectively. Unlike other images, angiograms have several characteristics which...[
Read more ]
Recently, medical image visualization has become a hot research topic and many novel ideas have been proposed for different practical applications. The general awareness of health and the continuously increasing number of patients suffering from various kinds of diseases show the importance of this research. Based on the observation of the severity of vascular diseases, we dedicate our effort to the research topic of vas-cular image visualization. We believe that effective visualization of images is critical to physicians for proper analysis and diagnosis of existing pathologies. Different from previous work related to medical visualization, we mainly focus on how to visualize angiograms more accurately and effectively. Unlike other images, angiograms have several characteristics which make them difficult to visualize using traditional meth-ods. In this thesis, we propose several methods to address the related issues and deliver a system to demonstrate our techniques.
Direct volume rendering is an effective way to visualize 3D vascular images for diagnosis of different vascular pathologies and planning of surgical treatments. An-giograms are typically noisy, fuzzy, and contain thin vessel structures. Therefore, some kind of enhancements are usually needed before direct volume rendering can begin. However, without visualizing the 3D structures in angiograms, users may find it difficult to select appropriate parameters and assess the effectiveness of the enhance-ment results. Also, traditional enhancement techniques cannot easily separate vessel voxels from other contextual structures with the same or very similar intensity. In this thesis, we propose a framework to integrate enhancement and direct volume rendering into one visualization pipeline using multi-dimensional transfer function tailored for visualizing the curvilinear and line structures in angiograms. The visualization result depends on viewers concern and therefore a proper user selection is a critical starting point for conveying results in an expected way. To ease the difficulties in defining the ROIs or specific vessels for visualization, we propose a MIP-guided method for user selection. Besides, as we found that rendering of small vessels is problematic using conventional approaches, we present a feature-preserving interpolation method to ren-der very thin vessels which are usually missed in traditional approaches. Furthermore, in order to increase the effectiveness and illustrative power of visualization, we in-troduce several non-photorealistic rendering methods into our system. Our goal is to effectively convey essential information about an image by presenting the structures in different manners using different rendering styles. In addition, we propose a view selection method for angiographic volume to ease the difficulty in finding meaningful views which can deliver useful information in the visualization process.
Post a Comment