Streamlines for Illustrative Real‐Time Rendering

Abstract

Line drawing techniques are important methods to illustrate shapes. Existing feature line methods, e.g., suggestive contours, apparent ridges, or photic extremum lines, solely determine salient regions and illustrate them with separate lines. Hatching methods convey the shape by drawing a wealth of lines on the whole surface. Both approaches are often not sufficient for a faithful visualization of organic surface models, e.g., in biology or medicine. In this paper, we present a novel object-space line drawing algorithm that conveys the shape of such surface models in real-time. Our approach employs contour-and feature-based illustrative streamlines to convey surface shape (ConFIS). For every triangle, precise streamlines are calculated on the surface with a given curvature vector field. Salient regions are detected by determining maxima and minima of a scalar field. Compared with existing feature lines and hatching methods, ConFIS uses the advantages of both categories in an effective and flexible manner. We demonstrate this with different anatomical and artificial surface models. In addition, we conducted a qualitative evaluation of our technique to compare our results with exemplary feature line and hatching methods.

Line Integral Convolution for Real‐Time Illustration of Molecular Surface Shape and Salient Regions

“…This work is motivated by molecular visualization, but not restricted to it. Furthermore, our approach provides a frame‐coherent illustration and is more tessellation‐independent in comparison to ConFIS [LMP13].…”

“…The second task was about a comparison with established line drawing techniques shown in Figure 6. The goal was to figure out if the new approach can keep up with the high‐quality hatching [ZISS04] and ConFIS [LMP13]. All participants stated that all three line drawing techniques were able to give a spatial impression of the model.…”

“…As the whole surface is covered with lines, relevant features may not be distinguished from unimportant ones. The recently presented ConFIS method [LMP13] tries to solve these issues. However, this method strongly depends on the underlying tessellation and the feature detection is based on the mean curvature.…”

Visualization and Extraction of Carvings for Heritage Conservation

“…While the method performs very well on artificial data, the processing of 3D-scans requires manual segmentation of the digital object and adjustment of parameters (see Sections 6.2.1 and 6.2.2 in [30]). A technique to illustrate surface features and enhance the spatial impression with lines was presented by Lawonn et al [21]. For a recent overview on the extraction of feature lines, we refer to [22].…”

Glyph‐Based Comparative Stress Tensor Visualization in Cerebral Aneurysms

“…We follow the idea by Lawonn et al [LLPH15], who generated quads along the streamlines to illustrate information. The streamline generation is based on the ConFIS method [LMP13]. For every generated quad, we assign coordinates which represent its width (orthogonal to the first PSD) in the range [−1,1].…”