Naoko Sawada

Naoko Sawada


Affiliation Faculty of Science and Technology — 2nd year in master degree
Team 疎性モデリング(Sparse Modeling, SpM)
Research theme Visualization of time-dependent, multi-dimensional blazer datasets
Hobbies Japanese archery, ski, drawing, beverage and delicious food
Special skill Identify the gender of cats from their face
E-mail @

Self-presentation


I love cats 🙂
I myself have a male cat in my house and my aunt has three male cats and a female cat.Who I respect the most are Hayao Miyazaki and Isao Takahata, who are Japanese animation film director and co-founder of Studio Ghibli.
I respect them for their extensive knowledge and attitude of pursuing new expressions in the hand-drawn animation.
My favorite movies are My Neighbor Totoro (by H. Miyazaki) The Tail of the Princess Kaguya (by I. Takahata).

About my research: TimeTubes


1. Overview of TimeTubes
Now I am studying on the visualization of observed blazar datasets. Astronomers have been observing blazars to solve the mystery
of the relativistic jet ejected from the central black hole of an active galactic nucleus. By using our visualization tool, TimeTubes, allows astronomers to analyze the dynamic behavior of and relationship among the multiple time-varying variables geometrically by visualizing a dataset as a 3D volumetric tube, as shown in Fig. 1 (a).

Fig. 1 User interface of TimeTubes.

The visualization space in Fig. 1 (a) shows the visualization result of a dataset, in which users are allowed to manipulate the tube interactively. Fig. 1 (b) can show the scatterplots between two arbitrary variables, which is mutually linked with TimeTubes view in Fig. 1 (a). The menu shown in Fig. 1 (c) allows for camera control, switching the display of the cruciform axes and placing a glyph that expresses the contour of the tube at each data sample. The operation panel shown in Fig. 1 (d) enables users to perform their visual analysis through overviewing, zooming, €filtering, details-on-demand, and relating with scatterplots.
Now we are trying to fuse multiple datasets from different observatories for the same blazar to help astronomers analyze the datasets more accurately and effectively. Visual data fusion and similarity search are the main functionalities of TimeTubes for the astronomers’ more accurate and effective data analysis.
2. Visual data fusion

Fig. 2 Categorized functions of visual data fusion with TimeTubes.

Blazar datasets includes uncertainties caused by errors and missing periods of observation. The current visual data fusion is designed to fuse observed datasets from two different observatories, Hiroshima University and University of Arizona, in order to ameliorate these kinds of uncertainties visually. We will introduce 2 modes and 4 options of visual data fusion, as summarized in Fig. 2. The Merge mode can ameliorate uncertainties caused by both errors and missing periods by increasing the amount of observed data used to form a single tube, whereas the Juxtaposition mode can ameliorate uncertainties caused by errors by comparing multiple datasets from different observatories.
Details of visual data fusion are discussed in our previous work [4].
3. Similarity search
Now we are trying to automatically detect important events or drastic changes, such like anomalies, flares of blazar intensity, rotation of observed polarization values, from long-term observed datasets. This functionality may lead to astronomers’ non-biased and effective data analysis. How we extract these important features is discussed in [9] in Japanese.

Achievements


[1] Naoko Sawada, Masanori Nakayama, Makoto Uemura, Hsiang-Yun Wu, and Issei Fujishiro. “TimeTubes: Visual Fusion for Ameliorating Uncertainty of Blazar Datasets from Different Observatories,” in Proceedings of the 79th IPSJ National Convention, Vol. 79, No. 4, pp. 85―86, Nagoya, Japan, March 2017, Student Encouragement Award (in Japanese).
[2] Naoko Sawada, Masanori Nakayama, Hsiang-Yun Wu, Makoto Uemura, and Issei Fujishiro. “TimeTubes: Visual Fusion for Ameliorating Uncertainty of Blazar Datasets from Different Observatories,” Refereed posters, in Proceedings of 10th IEEE Pacific Visualization Symposium (PacificVis 2017), pp. 336―337, Seoul, Korea, April 2017.
[3] Issei Fujishiro, Shigeo Takahashi, Kazuho Watanabe, Hsiang-Yun Wu, Naoko Sawada, and Makoto Uemura. “On the Perspicuity of Multidimensional Data Visualization,” Poster, in the 1st Public Symposium in the 2017 business year of KAKENHI Grant-in-Aid for Scientific Research on Innovative Areas “Initiative for High-Dimensional Data-driven Science through Deepening of Sparse Modeling”, June 2017 (in Japanese).
[4] Naoko Sawada, Masanori Nakayama, Hsiang-Yun Wu, Makoto Uemura, and Issei Fujishiro. “TimeTubes: Visual Fusion and Validation for Ameliorating Uncertainty of Blazar Datasets from Different Observatories,” Short papers, in Proceedings of the Computer Graphics International Conference (CGI 2017), Article No. 14, Yokohama, Japan, DOI: 10.1145/3095140.3095154, June 2017. [ACM Digital Library]
[5] Makoto Uemura, Ryosuke Itoh, Ioannis Liodakis, Dmitry Blinov, Masanori Nakayama, Longyin Xu, Naoko Sawada, Hsiang-Yun Wu, Issei Fujishiro. “Optical polarization variations in the blazar PKS 1749+096,” Publications of the Astronomical Society of Japan (PASJ), Vol. 69, No. 6, Article No. 96, DOI: 10.1093/pasj/psx111, November 2017. [arXiv][Oxford Academic]
[6] Naoko Sawada, Masanori Nakayama, Hsiang-Yun Wu, Makoto Uemura, and Issei Fujishiro. “TimeTubes: Visual Fusion for Detailed and Precise Analysis of Time-Varying Multi-Dimensional Datasets,” in the International Meeting on “High-Dimensional Data-Driven Science” (HD3-2017), September 2017.
[7] Issei Fujishiro, Naoko Sawada, Masanori Nakayama, Hsiang-Yun Wu, Kazuho Watanabe, Shigeo Takahashi, and Makoto Uemura. “TimeTubes: Visual Exploration of Observed Blazar Datasets,” to appear in Journal of Physics: Conference Series (JPCS), Kyoto, Japan, 2018.
[8] Issei Fujishiro, Shigeo Takahashi, Kazuho Watanabe, Hsiang-Yun Wu, Naoko Sawada, Makoto Uemura. “Consolidation of Visualization Platform Toward Facilitating Sparse Modeling,” Poster, in the 5th Public Symposium of KAKENHI Grant-in-Aid for Scientific Research on Innovative Areas “Initiative for High-Dimensional Data-driven Science through Deepening of Sparse Modeling”, December 2017 (in Japanese).
[9] Naoko Sawada, Masanori Nakayama, Makoto Uemura, and Issei Fujishiro. “TimeTubes: Feature Extraction of Observed Blazar Datasets for Detailed and Efficient Data Analysis,” in the 284th Reports of the Technical Conference of The Institute of Image Electronics Engineers of Japan, pp. 19―23, Hiroshima, Japan, March 2018 (in Japanese).