Creating Character Networks with Kinship Relations from Comics

  • Harumi Murakami Osaka City University
  • Yusuke Nagaoka Osaka City University
  • Ryota Kyogoku Osaka City University


To find and understand comic contents, we present an approach of
creating character networks with semantic relations from comics using
the panels and speech balloons. First, we build a dataset that
contains a comic's characters and its words in speech balloons
separated by the panels. Second, we extract the semantic relations
among characters using the words in the speech balloons and
panels. Third, we develop character networks by identifying the
important characters using the frequencies of characters and their
relatedness, and then we add semantic relations to these networks. As
an initial step in finding semantic relations, we focus on kinship
relations. We implement our approach and successfully build character
networks from four popular Japanese comics. Finally, we evaluate our
kinship relation extraction algorithms and the created character


W. S. Wang, “Globalizing manga: From Japan to Hong Kong and beyond,”

Mechademia, pp. 23–45, 2006.

J. J. Miller, “The last million-selling comic book in North America?

It’s Batman vs. Pokemon for the title,” 2014. [Online]. Available:

Weekly Shonen Jump. Shueisha (in Japanese), 1993, vol. 43.

Weekly Shonen Jump. Shueisha (in Japanese), 2014, vol. 50.

H. Murakami, R. Kyogoku, and H. Ueda, “Creating character connections from manga,” in Proc. 3rd Int. Conf. Agents and Artificial Intelligence (ICAART 2011), vol. 1, 2011, pp. 677–680.

A. Toriyama, Dragon Ball. Shueisha (in Japanese), 1992, vol. 32.

Y. Nagaoka, R. Kyogoku, and H. Murakami, “Creating character connections from comics using frames and words in balloons,” in Human Communication Group Symposium 2015, 2015, pp. 12–13, (in Japanese).

H. Murakami, Y. Nagaoka, and R. Kyogoku, “Creating character networks from comics using frames and words in balloons,” in Proc. 2018 7th IIAI Int. Congr. Advanced Applied Informatics (IIAI-AAI 2018), 2018, pp. 1–6.

T. Ueno and K. Kazama, “Method of creating character networks based on speech timing in comics,” in Proc. 11th ARG SIG-WI2, 2017, pp. 1–6, (in Japanese).

D. K. Elson, N. Dames, and K. R. McKeown, “Extracting social networks from literary fiction,” in Proc. 48th Annu. Meeting Association for Computational Linguistics (ACL 2010), 2010, pp. 138–147.

H. Nishihara, “Extraction of relations among characters in literary texts,” Master’s Thesis, JAIST, 2015, (in Japanese).

D. Kokkinakis and M. Malm, “Character profiling in 19th century fiction,” in Proc. Language Technologies for Digital Humanities and Cultural Heritage Workshop, 2011, pp. 70–77.

D. Kokkinakis, M. Malm, J. Bergenmar, and A. Ighe, “Semantics in storytelling in Swedish fiction,” in DATeCH ’14 Proc. 1st Int. Conf. on Digital Access to Textual Cultural Heritage 2014, 2014, pp. 137–142.

J. Lee and C. Y. Yeung, “Extracting networks of people and places from literary texts,” in Proc. 26th Pacific Asia Conf. Language, Information, and Computation (PACLIC 2012), 2012, pp. 209–218.

J. Goto, Y. Yagi, A. Aizawa, and S. Sekine, “Generation of correlation charts from tv programs based on anaphora resolution,” in 22nd Annu. Conf. Japanese Society for Artificial Intelligence, 2008, (in Japanese).

A. Ramakrishna, V. Martinez, N. Malandrakis, K. Singla, and S. Narayanan, “Linguistic analysis of differences in portrayal of movie characters,” in Proc. 55th Annu. Meeting Association for Computational Linguistics (ACL 2017), 2017, pp. 1669–1678.

Y. Matsuo, J. Mori, M. Hamasaki, K. Ishida, T. Nishimura, H. Takeda, K. Hashida, and M. Ishiduka, “Polyphonet: An advanced social network extraction system,” in Proc. WWW2006, 2006, pp. 397–406.

S.-B. Park, Y.-W. Kim, M. Uddin, and G.-S. Jo, “Character-Net: Character network analysis from video,” in WI-IAT ’09 Proc. 2009 IEEE/WIC/ACM Int. Joint Conf. Web Intelligence and Intelligent Agent Technology, vol. 1, 2005, pp. 305–308.

K. Yuan, R. Ji, H. Yao, X. Sun, P. Xu, and X. Liu, “VisualCor System: search actor correlations in tv series,” in ICIMCS ’09: Proc. 1st Int. Conf. Internet Multimedia Computing and Service, 2009, pp. 213–218.

M. Matsushita, “The possibility of comic computing,” in Proc. 2nd ARG WI2, no. 15, 2013, (in Japanese).

M. Matsushita, “The past and future of comic computing,” in Proc. 11th JSAI Interactive Information Access and Visual Mining, 2015, pp. 12–19, (in Japanese).

X. Pang, Y. Cao, R. W. H. Lau, and A. B. Chan, “A robust panel extraction method for manga,” in MM ’14 Proc. 22nd ACM int. conf. Multimedia, 2014, pp. 1125–1128.

Q. Guo, K. Kato, N. Sato, and Y. Hoshino, “An algorithm for extracting text strings from comic strips,” in SIGGRAPH ’06 ACM SIGGRAPH 2006 Research posters, no. 76, 2006.

W.-T. Chu and W.-W. li, “Manga FaceNet: Face detection in manga based on deep neural network,” in ICMR ’17 Proc. 2017 ACM Int. Conf. Multimedia Retrieval, no. 76, 2017, pp. 412–415.

A. Fujimoto, T. Ogawa, K. Yamamoto, Y. Matsui, T. Yamasaki, and K. Aizawa, “Manga109 dataset and creation of metadata,” in MANPU ’16 Proc. 1st Int. Workshop on coMics Analysis, Processing and Understanding, no. 2, 2016.

M. Kishimoto, Naruto. Shueisha (in Japanese), 2009, vol. 46.

M. Kishimoto, Naruto. Shueisha (in Japanese), 2010, vol. 50.

A. Toriyama, Dragon Ball (Kanzenban). Shueisha (in Japanese), 2004, vol. 29.

A. Toriyama, Dragon Ball (Kanzenban). Shueisha (in Japanese), 2004, vol. 34.

Y. Kamio, Boys Over Flowers. Shueisha (in Japanese), 1997, vol. 16.

Y. Kamio, Boys Over Flowers. Shueisha (in Japanese), 1998, vol. 20.

C. Koyama, Space Brothers. Kodansha (in Japanese), 2008, vol. 1.

C. Koyama, Space Brothers. Kodansha (in Japanese), 2009, vol. 5.

Technical Papers