It is difficult for beginners to create well-balanced etched latte art patterns using two fluids with different viscosities, such as foamed milk and syrup. However, it is not easy to create well-balanced etched latte art even while watching process videos that show procedures. In this paper, we propose a system that supports beginners in creating well-balanced etched latte art by projecting the etching procedure directly onto a cappuccino. In addition, we examine the similarity between etched latte art and design templates using background subtraction. The experimental results show the progress in creating well-balanced etched latte art using our system.

Effective 3D modeling for urban landscape reconstruction is one of the most important fields in 3D Computer Graphics. There have been proposed many modeling techniques. Procedural modeling method [Müller et al. 2006] is powerful and effective, however, it reconstructs buildings one by one precisely. Each building has to be specified as a procedure of various kinds of construction parts of buildings. Therefore, it takes long time to reconstruct urban landscapes.

In painting pictures, painters sometimes use tools such as chopsticks and clothes, which are not as familiar as brush and pencil, solely or in combination with other tools [1][2](Fig. 1). This paper presents on a painting tool, PETICA (Fig. 2), which we developed to achieve our aim to reproduce strokes represented by aforementioned various tools.

We propose a new modeling method for the appearance of fallen snow formed by snow drift near buildings and other obstacles. Instead of computational fluid dynamics simulation for wind-driven fallen snow, we employ the Pre-computed Radiance Transfer (PRT) technique [Sloan et al. 2002] to handle snowflakes which fly, hit and fall with the wind blowing to all directions around obstacles. This approach approximates snow accumulation with snow drift without heavy fluid dynamics simulation computation.

We have been restoring city landscape of Edo in the Edo era, based on archaeological documents, with 3DCG. Edo is old name of Tokyo. We call the city landscape restored with 3DCG as cyberspace hereafter. As restoration methods of cyberscapes, several procedural models[1][2] have been proposed, by which buildings with stones and bricks were generated. However, it is difficult to apply them to generate wooden buildings without complete drawings and vulnerable to weathering.

We propose an intermediary approach to utilizing existing Web resources as educational material. The Web is attractive for educational use but it is not completely suitable. To fix the remedy, we developed an integrated proxy server, called WebAngel, which can automatically and individually personalize Web resources for students. We present the mechanism of WebAngel to support learning experiences on the Web and extends the educational possibilities of Internet usage.

In artistic design education, the teacher instructs each student individually face to face. As a result, it is difficult to share coaching with a third party or to teach distantly. To solve these problems, we have developed Manga University, which is a web-based application to aid artistic design education. It enables distant teaching or shared teaching and provides a learning portfolio for collaboration. It is applicable to several other types of artistic design education, such as fashion design or GUI design for software or the web.

This paper describes a fast analytic algorithm that generates exact highlights and soft shadows from area light sources. In order to realize fast shadowing, we propose the ray‐oriented buffer which segments 3D space by following light rays from polygonal sources. Each cell of the buffer stores objects that intersect a related subspace. Candidate objects which may cast shadows onto a point are selected by referring to the buffer. The candidates are then tested with their shadow bounding volumes to suppress objects that never occlude light sources. In addition, we propose the cross scanline clipping algorithm. It quickly determines the exact regions of uncovered area light sources with simple silhouette generation. Both diffuse and specular reflections are computed by integrating light rays from the uncovered sources. Experimental results confirm the high performance of the proposed method.

Abstract This paper presents a fast shadowing algorithm for linear light sources that uses a ray‐oriented buffer. Space segmentation by the buffer guarantees that if a point is included in a subspace, all light rays toward the point are also contained in the subspace. Each cell of the buffer stores a list of objects that lie within or intersect the subspace allocated to the cell. Therefore, candidate objects, those that may cast shadows onto a point, are determined by referring to the cell where the point is mapped. In addition, whether each candidate object actually casts shadows or not is tested with the bounding‐volume of the shadow space to reduce the number of objects subjected to expensive light clipping. The bounding‐volumes are also stored in the buffer. For efficiently generating the ray‐oriented buffer, we present the cylindrical scan‐conversion algorithm. The algorithm preconverts objects' surfaces to trapezia to decrease the light clipping cost, then connects the trapezia to the buffer cells. Due to the above improvements, our algorithm achieves over 10 times faster shadow generation compared to the conventional methods. Experimental results confirm that our method can generate realistic images with soft shadows in a few minutes.

Abstract We propose the Cross Scanline Buffer which preserves the result of hidden surface removal as performed by the Cross Scanline Algorithm. The Cross Scan Buffer reduces image re‐generation time and eliminates aliasing artifacts even if the image is arbitrarily scaled. Perfect anti‐aliasing is achieved because the Cross Scanline Algorithm analytically determines visible polygonal surfaces and divides them into sets of triangles and trapezia. The Cross Scan Buffer supports the various applications that currently use the conventional buffering methods for anti‐aliasing. This paper introduces and tests three applications: image scaling, shadow creation, and texture mapping. Experimental results verify that the Cross Scan Buffer is very powerful yet efficient.

This paper derives a shading model for area light sources which covers both diffuse and specular reflection. The shading model assumes ideally diffuse polygonal light sources and uses Phong’s reflection model. The model can accurately integrate the intensities of diffuse and specular reflection without simulating an area light source as an array of point light sources. To simplify the reflection integration, each light source is transformed from the Cartesian coordinate system into the polar system. The light source is projected onto a unit sphere and then triangulated along great circles of the unit sphere. Finally, the integration value is calculated by polynomial approximation. Since our method can accurately integrate both diffuse and specular reflection, it can generate images that are more photorealistic than conventional methods. Because point light sources are not employed, highlight roughness is completely suppressed. Several images are presented that show the advantages of our method.

The G-buffer method is applied to NC machining. A total NC system is created that consists of all essential functions, such as tool path generation, path verification, and feed rate control. Moreover, any combination of object surface and tool shape is acceptable. By utilizing G-buffers created from a parallel projection, the required NC functions are realized as image processing operations. This ensures that the NC software is independent from surface description. Conventional rendering software can be used to make the G-buffers. Any surface can be milled if it can be rendered by parallel projection. Tool shape changes can be easily handled by changing the image processing filters. 3D examples of geometric surfaces, mesh data, and volume data are milled with this method, and the results show that the method is very effective.

This paper proposes a new hidden surface removal algorithm which is based on the scanline algorithm but scans in two directions, horizontally and vertically. Named the cross scanline algorithm, it can efficiently detect all polygons and calculate their exact projected areas in each pixel even if the polygons are much smaller than the pixel. Comparisons with the regular sub-scanlines algorithm show that high quality anti-aliased images can be generated.

We propose a new rendering technique that produces 3-D images with enhanced visual comprehensibility. Shape features can be readily understood if certain geometric properties are enhanced. To achieve this, we develop drawing algorithms for discontinuities, edges, contour lines, and curved hatching. All of them are realized with 2-D image processing operations instead of line tracking processes, so that they can be efficiently combined with conventional surface rendering algorithms.Data about the geometric properties of the surfaces are preserved as Geometric Buffers (G-buffers). Each G-buffer contains one geometric property such as the depth or the normal vector of each pixel. By using G-buffers as intermediate results, artificial enhancement processes are separated from geometric processes (projection and hidden surface removal) and physical processes (shading and texture mapping), and performed as postprocesses. This permits a user to rapidly examine various combinations of enhancement techniques without excessive recomputation, and easily obtain the most comprehensible image.Our method can be widely applied for various purposes. Several of these, edge enhancement, line drawing illustrations, topographical maps, medical imaging, and surface analysis, are presented in this paper.

Pencil tracing, a new approach to ray tracing, is introduced for faster image synthesis with more physical fidelity. The paraxial approximation theory for efficiently tracing a pencil of rays is described and analysis of its errors is conducted to insure the accuracy required for pencil tracing. The paraxial approimation is formulated from a 4x4 matrix (a system matrix) that provides the basis for pencil tracing and a variety of ray tracing techniques, such as beam tracing, ray tracing with cones, ray-object intersection tolerance, and a lighting model for reflection and refraction. In the error analysis, functions that estimate approximation errors and determine a constraint on the spread angle of a pencil are given.The theory results in the following fast ray tracing algorithms; ray tracing using a system matrix, ray interpolation, and extended 'beam tracing' using a 'generalized perspective transform'. Some experiments are described to show their advantages. A lighting model is also developed to calculate the illuminance for refracted and reflected light.

Abstract The. architecture of, and a performance estimate for, the dedicated graphics processor ‐ SIGHT, a high‐speed ray tracing machine, are presented. Ray tracing is a vector calculation in 3–dimensional (3–D) space and each ray to be traced can be calculated independently. The SIGHT is designed so that it speeds up the execution of the ray tracing algorithm at two levels of parallel operations: the pixel‐level parallel operation by multiprocessors and the instruction‐level parallel operation by the new processing element (PE) architecture. The features of the PE are: