論文

Motion retargeting is a well-known method for generating character animations in computer graphics, where the original motion of an actor can be transferred to an avatar with a different skeletal structure and/or limbs' length. A specific type of retargeting for manipulating various avatars interactively is referred to as puppetry. However, previous learning-based works of puppetry cannot be said to be fully interactive, or require a large amount of metadata about the avatar's skeleton as well as training data of motions. Thus, we attempted to integrate existing methods by taking into account skeletal similarity between an actor and his/her avatar, to come up with an interactive and intuitive motion retargeting method that only requires a relatively small amount of training animations together with simple input metadata. Moreover, by classifying avatar's body parts in a procedural manner, a more flexible motion puppetry was realized, where the system user is allowed to specify desirable part-to-part correspondences.

We proposed a stable and efficient particle-based method for simulating highly viscous fluids that can generate coiling and buckling phenomena and handle variable viscosity. In contrast to previous methods that use explicit integration, our method uses an implicit formulation to improve the robustness of viscosity integration, therefore enabling use of larger time steps and higher viscosities. We use Smoothed Particle Hydrodynamics to solve the full form of viscosity, constructing a sparse linear system with a symmetric positive definite matrix, while exploiting the variational principle that automatically enforces the boundary condition on free surfaces. We also proposed a new method for extracting coefficients of the matrix contributed by second-ring neighbor particles to efficiently solve the linear system using a conjugate gradient solver. Several examples demonstrated the robustness and efficiency of our implicit formulation over previous methods and illustrated the versatility of our method.

Exploring feature subspaces is one of promising approaches to analyzing and understanding the important patterns in multivariate data. If relying too much on effective enhancements in manual interventions, the associated results depend heavily on the knowledge and skills of users performing the data analysis. This paper presented a novel approach to extracting feature subspaces from multivariate data by incorporating biclustering techniques. The approach has been maximally automated in the sense that highly-correlated dimensions are automatically grouped to form subspaces, which effectively supports further exploration of them. A key idea behind our approach lies in a new mathematical formulation of asymmetric biclustering, by combining spherical k-means clustering for grouping highly-correlated dimensions, together with ordinary kmeans clustering for identifying subsets of data samples. Lower dimensional representations of data in feature subspaces are successfully visualized by parallel coordinate plot, where we project the data samples of correlated dimensions to one composite axis through dimensionality reduction schemes. Several experimental results of our data analysis together with discussions were provided to assess the capability of our approach.

本論文では，3Dスキャナから採取した3Dボディデータを分析することで，スカートの高精度な設計から，その展開に基づいたスカート原型作製までカバーするデザインシステムを提案する．本システムは，（1）人体のポリゴンデータから意図するデザインに沿った正確な基準断面データ生成可能，（2）基準断面データを用いた立体裁断方式による自動作図機能，（3）多くの体型に対応可能なパターン生成という特長をもつ．これにより，個々の3Dボディデータが高速に分析でき，体型にフィットした着心地の良いスカート原型を短時間に得ることが可能となった．

We report on the variation in the optical polarization of the blazar PKS 1749+096 observed in 2008–2015. The degree of polarization (PD) tends to increase in short flares having a time-scale of a few days. The object favors a polarization angle (PA) of 40◦–50◦ at the flare maxima, which is close to the position angle of the jet (20◦–40◦). Three clear polarization rotations were detected in the negative PA direction associated with flares. In addition, a rapid and large decrease in the PA was observed in the other two flares, while another two flares showed no large PA variation. The light curve maxima of the flares possibly tend to lag behind the PD maxima and color-index minima. The PA became -50◦ to -20◦ in the decay phase of active states, which is almost perpendicular to the jet position angle. We propose a scenario to explain these observational features, where transverse shocks propagate along curved trajectories. The favored PA at the flare maxima suggests that the observed variations were governed by the variations in the Doppler factor, $delta$. Based on this scenario, the minimum viewing angle of the source, min = 4:8◦–6:6◦, and the location of the source, Δr> 0.1 pc, from the central black hole were estimated. In addition, the acceleration of electrons by the shock and synchrotron cooling would have a time-scale similar to that of the change in $delta$. The combined effect of the variation in $delta$ and acceleration/cooling of electrons is probably responsible for the observed diversity of the polarization variations in the flares.

To visualize the behavior of complex functions, we need four dimensions. We raise the question that has been neglected: "What does the Riemann sphere's axis stand for?" The answer can be obtained by setting the immersive environment at the sphere's origin, which is always undefined in projective geometry.

本研究では，初心者でも直感的にデフォルメされた2D グラフィックデザインを生成できるシステム|Swellart を提案する．我々は，システムを開発するうえで, 変形の方法のなかでも膨張に着目している. ユーザが描画した枠の形状に合わせて対象の画像が膨張する過程を単純に示すことで, すべてのユーザが容易に変形過程やその結果を想像させられる．膨張方法として，まずユーザが画像を指定して，デフォルメしたい範囲をユーザがくり抜き，その範囲内のメッシュ分割を行う．分割され た各セルに対し，さまざまな制約をユーザが指定して膨張を行うことで，多様なデフォルメデザインを生成できる．本論文では，Swellart の処理手順と結果を示し，今後の方向性を議論する．

コンピュータグラフィックス（CG）において糸を細部まで忠実に再現することは，困難な課題として知られている．糸は織物や編み物，衣服を構成するために必要であり，現実世界を仮想空間に再現するために不可欠な素材である．これまでに開発された糸の生成手法は，ボリュームベースと繊維ベースの二種類に分類できる．近年提案された繊維ベース手法では，糸の断面をCT スキャンし，その細部を精密に表現している．近年の先進的な手法により，CG における糸の表現力は劇的に向上し，なおかつその生成に必要なコストも減少した．しかし，糸の周囲に生える毛羽立ちの生成プロセスは，未だに挑戦的な課題である．特に，表面の毛羽立ちが独特な形状となるパイル織物は，タオルや車のシートなど生活のあらゆる場所で使用される一般的な素材であるにも関わらず，既存のフレームワークでは表現が不可能である．加えて，上述した繊維ベース手法では糸の生成プロセスに注目していたが，その糸を使用した織物の生成フレームワークについては，筆者らの知る限り，未だに提案されていない．これらの課題を解決するため，本論文では既存のフレームワークを拡張することでパイル繊維をもつ糸の生成に対応するとともに，この糸から自動的にパイル織物を生成する手法を提案する．

As is well-known, in quantum mechanics the complex wave plays an essential role, which can explain the natural phenomena very well. In other words, it can be said that the rules of nature are written with logic using complex numbers. We are accustomed to the convenience of visualizing the overall behavior of a function by means of its graph. In the case of a complex function, however, this approach has not seemed workable for a long time because to depict a pair of complex numbers (an input z and the output of a function f) we would need four dimensions. Especially in mathematical visualization, we have demonstrated thus far that how the fourth dimension is conceptualized seems to be successfully justified by using a cross section and a projection of an object appropriately, described in terms of polar coordinates (instead of Cartesian coordinates). At the end of this paper, we raise the question that has been neglected thus far: “What does the Riemann sphere’s axis stand for?” We show that the answer can be obtained only by observing from the inside the sphere by setting the viewpoint of the immersive environment on the point (0, 0), which is always undefined in projective geometry.

Astronomers have been observing blazars to solve the mystery of the relativistic jet. A technique called TimeTubes uses a 3D volumetric tube to visualize the time-dependent multivariate observed datasets and allows astronomers to interactively analyze the dynamic behavior of and relationship among those variables. However, the observed datasets themselves exhibit uncertainty due to their errors and missing periods, whereas periods interpolated by TimeTubes result in a di‚erent type of uncertainty. In this paper, we present a technique for ameliorating such data- and mappinginherent uncertainties: visual fusion of datasets for the same blazar from two different observatories. Visual data fusion with Time-Tubes enables astronomers to validate the datasets in a meticulous manner. This is a short paper presented at CGI2017, held at Keio University Hiyoshi Campus, Yokohama, Japan, June 27-30 (presentation day: June 28), 2017.

本研究では，管楽器を演奏するキャラクタの吹奏アニメーションを，音楽に容易に同期させるため，音源から自動的に運指のアニメーションを生成することを目指す．より具体的には，ウインドシンセサイザを用いて作成したMIDI 音源を解析することにより，吹奏の情報を得る．得られた情報をキャラクタの指や腕，および管楽器のピストンやスライドに適用することにより，運指が音源に同期した，自然な吹奏アニメーションを実現する．ポスタNo.41として発表．

本発表では，2 次元平面上に縦横の方向性をもたせた繊維を配置し，繊維同士をつなげて繊維のシート（紙繊維シート）を生成する．さらにシートにはシートを充填する填料，繊維の回転に反発するヒンジばねのモデルを組み込み，シートに力が加えられたときこれらに発生する力学的な作用を考慮して破断のシミュレーションを行う．ポスタNo.33として発表．

本発表では，骨格の解剖学的なデータと，腱，筋肉，皮膚，血管の人工的なデータを用いて，各器官の形状をスカラ関数として表現し，組み合わせる．このようにしてモデリングされた手の準解剖学的なモデルを動かし，スカラ関数を変化させることで，内部構造の動きによる手の表面の変化を陰的にモデリングできる手法を提案する．ポスタNo.28として発表．

2017年6月17-18日にお茶の水女子大学（東京都文京区）で開催された音学シンポジウム2017（第115回 音楽情報科学研究会）のポスタ発表．

Datasets obtained through recently advanced measurement techniques tend to possess a large number of dimensions. This leads to explosively increasing computation costs for analyzing such datasets, thus making formulation and verification of scientific hypotheses very difficult. Therefore, an efficient approach to identifying feature subspaces of target datasets, that is, the subspaces of dimension variables or subsets of the data samples, is required to describe the essence hidden in the original dataset. This paper proposes a visual data mining framework for supporting semiautomatic data analysis that builds upon asymmetric biclustering to explore highly correlated feature subspaces. For this purpose, a variant of parallel coordinate plots, many-to-many parallel coordinate plots, is extended to visually assist appropriate selections of feature subspaces as well as to avoid intrinsic visual clutter. In this framework, biclustering is applied to dimension variables and data samples of the dataset simultaneously and asymmetrically. A set of variable axes are projected to a single composite axis while data samples between two consecutive variable axes are bundled using polygonal strips. This makes the visualization method scalable and enables it to play a key role in the framework. The effectiveness of the proposed framework has been empirically proven, and it is remarkably useful for many-to-many parallel coordinate plots.

Blazars are active galactic nuclei whose relativistic jets ejected from the central black hole are pointing toward the Earth. Astronomers have attempted to classify blazars, whereas it is difficult to analyze the time-dependent multivariate datasets with the conventional visualization methods, such as animeted scatterplot matrices. In our previous study, a new visualization scheme, called TimeTubes, was proposed, which allows the astronomers to analyze dynamic changes in and feature causality among the multiple time-varying variables. In this poster, we present a core idea to ameliorate data-inherent and mapping-inherent uncertainty through visual fusion of datasets for the same blazar from two different observatories.