Katsuhiro Kitajima

Computer animation of swaying trees based on physical simulation

Abstract This paper presents a series of techniques for generating animations of trees swaying in the wind, in consideration of the influences that the tree shapes and leaf sizes give to the air current. To do the simulation of the wind around a tree having a complicated shape, it is necessary to consider the influence that some objects obstructing the wind such as leaves or branches give. Generally, the following problem occurs when we use the incompressible Navier–Stokes equations in a physical simulation model of the wind. Computational complexity increases because of considering the details of tree shapes, so it is difficult to generate the animations in real-time. Therefore, this paper proposes a novel method that reduces the computational complexity and realizes an animation in real-time, by means of a boundary condition map expressing space distribution of resistances from tree models automatically. In this case, we treat the parts that have similar shapes such as leaves and branches as simple damping factors decreasing the wind velocity. And also, it has another advantage that the influences between a tree and others can be rapidly calculated by using a hierarchical calculation method. Finally, through many experiments using these methods, it is shown that real-time animations of swaying trees in the wind can be realized.

HIMADES-1: a hierarchical machine design system based on the structure model for a machine

Abstract A structure model is presented which can represent not only the simplified geometry of an individual part but also the whole structure of a machine. The model is derived from observation of the general features of a machine, ie the hierarchy and the topology (connectivity) among parts. The model plays an important role in developing a CAD system to aid designers in the primary design stage, because a general method for representing the whole structure of a machine is needed at that stage. An interactive machine design system named HIMADES-1 has been developed. The data structure of the system is based on the above structure model. The system has some original features, the most distinctive of which is the unified algorithm for dimension matching and motion analysis which reduces design labour in the primary stages of machine design.

A Proposal of Assembly Model Framework Specialized for Unified Parametrics

Almost all of the previous papers discussing parametrics problems have a common premise that the necessary information solving parametrics is given by a human (not transformed from CAD data) and much attention has been paid in the methods of solving them. Recently, the word ‘assembly model’ has spread widely. We have already proposed in another paper a method of unified parametrics including not only conventional parametrics problems for dimensions of a single part, but kinematics problems. In this paper, we discuss the relationship between our method of unified parametrics and the assembly model.

A shell‐oriented boolean set operations algorithm suited for the b‐reps based on boundary edge loops

This paper presents a shell-oriented Boolean set operations algorithm which is simple and makes the most of the F-L-E-V type B-reps based on boundary edge loops of a face. The characteristic feature of the algorithm is that it resolves an operation between solids into plural operations between faces, generates intermediate open shells and finally reconstructs a new solid by combining them. The method is quite different from the conventional edge-oriented algorithm, e.g., that of BUILD, which is suited for the well-known winged-edge type B-reps based on the information with respect to the individual edges. First, the outline of the conventional set operations algorithm is shown, which is suited for the winged-edge type B-reps and is used in almost all conventional solid modellers. Afterward, in comparison to it, our algorithm is discussed in detail. An algorithm for the case of a basic intersection, with which the basic algorithms in other set operations, i.e., union and set difference, are also realized, is described first and then some essential expansions are made to meet more complex and particular conditions.

Interactive robot simulator for high-level tasks

Abstract A robot simulator is developed that enables a robot operator to make plans of high-level tasks easily without knowing the details of the robot, before its off line teaching and programming. The simulator is based on the structural model developed in the CAD system Himades-1, which was devised for general assembly units with rotating and prismatic pairs. First, some unique basic functions that use the structural model are introduced. Then the general method, independent of the type of robot, is reported, which provides the consistent and intelligent algorithm to extract the necessary transformation matrices systematically from the structural model, to set up automatically the formulas for backward motion analysis (inverse kinematics) as non-linear simultaneous equations, and to solve them. Finally, some sophisticated simulations for high-level tasks are shown.

A method of simulating the fusion of speech and facial expression for individuals based on the GFFD method

This paper proposes a novel method of simulating the combination of speech and facial expression with anatomical basis, which can cope with any individual faces without any expensive measuring machines. In our previous paper, we presented a method of modeling muscle models around the mouth, especially the orbicularis oris model and the ones supporting it, which play important roles to decide the movement of skin around the mouth when we speak. The facial muscle models are represented as wire-frame models and their vertices play an important role when we treat them as operation points when applying our GFFD (Generalized Free Form Deformation) method. We experimentally decided the position of operation points of those muscle models when we utter the vowel /a/, /i/, /u/, /e/, /o/ and /n/ and consonants on the generic model. In this paper, we first report the improvement on the above method by considering jaw rotation, etc. Then, we present a new method to represent the fusion of the skin motions caused by both speech and facial expression. Some examples will finally be shown.

Study on a method of restoring the 3D virtual nail model from the 2D image data based on the GFFD deformation method

This paper proposes a novel method of generating a 3D virtual nail model from multiple camera image data based on the GFFD deformation method. 3D virtual nail models for individuals can be generated easily and rapidly with considerably low cost. The outline of the whole procedures is as follows. First, we take three pictures of an individuals nail. Second, we extract operation points of the nail in each image, which is arranged on the outline and surface of the nail. Third, we obtain the corresponding 3D positions of the operation points using a stereo computing method. Finally, we deform the 3D generic virtual nail model into an individuals one with GFFD based on the obtained 3D positions of the operation points. We made some experiments to evaluate whether the errors of the restored nail shape model is within the acceptable error range. By comparing the restored nail model with the model obtained by scanning the nail with a 3D digitizer, we evaluated the accuracy of the restored model. The result of the experiment shows that we can get individual nail models with sufficient accuracy.

Controlling color regions of leaves with painting techniques for landscape arts

In this paper, we propose a non-photorealistic rendering (NPR) framework for creating landscape art in traditional animation. The proposed method is based on painting techniques for landscape arts, which enables a user to reproduce features of the target work of art and its style. Landscape (background) arts is generally used in animation productions. One image of landscape art and multiple layers of moving characters compose a scene. Landscape art is one of most important factors for making the impression of an animation. We propose a NPR technique that reproduces painting techniques for landscape arts such as leaf painting based on just a few colors, shading to produce a spherical effect. The technique developed by the authors generates dynamic landscape art having a swaying motion in the wind by applying the animation generation technique for swaying trees to it.

Effective exercise instruction system using virtual human

This study aims at developing an aerobics exercise instruction software for personal use by introducing an instruction program best fit to the physical conditions of respective exerciser. The software uses the combination of virtual instructor (virtual human) and Karvonen method in order to develop a comprehensive personal aerobics exercise instruction software. Karvonen method is an algorithm of relationship between the movement of the virtual human and exercise intensity. In this study, its effectiveness is verified and confirmed.The study is mainly focused on the development of synchronization technique between the speed of animation figures movement for controlling exercise intensity and speed of the sound output, which is the most important technique in the application of the study result into practical use.

Gaussian-based Free Form Deformation and its Application to Fashion Design

This paper introduces an interactive and intuitive fashion design system using Gaussian-Based Free Form Deformations (GFFDs). GFFDs have several advantages over conventional Free Form Deformations. In GFFDs, Control points need not be positioned parallelpipedically and the locality of deformation can be easily controlled. Thus, GFFDs give designers more freedom in designing clothes. The theory of GFFDs, as well as some examples of clothes designed by our method, is presented.