A Study on Anxiety Reduction of Reader-dependent “Jirai” Expressions in Comics
>100 Views
October 10, 24
スライド概要
Since comics are diverse, there are some depictions that readers like and some that they do not like. If readers find a part of a comic that they dislike, they can continue reading it even if they can skip that part, but there is a possibility that they will stop reading it entirely if they actually read it. Therefore, we propose a method for allowing readers to enjoy comics without worrying about depictions that they do not like. To realize this method, we developed a system that allows readers to flag depictions they dislike while reading comics and conducted data collection experiments. We also implemented a system in which jirai (content to be avoided) flags and announcements are given to readers while reading a comic, and we examined the feasibility of the system by operating it for about four weeks. We confirmed that the flagging and jirai announcements were performed during the system operation. In addition, the evaluation of jirai judgments for comics using the Vision API showed that while there is potential for AI to make judgments, there are still difficulties in judging detailed depictions.
明治大学 総合数理学部 先端メディアサイエンス学科 中村聡史研究室
関連スライド
各ページのテキスト
A Study on Anxiety Reduction of Reader-dependent “Jirai” Expressions in Comics Yuki Nakagawa, Risa Ito, Satoshi Nakamura Meiji University, Japan
Introduction People have a sense of dislike or fear for various things. Blood Insect Trypophobia 2
Introduction People have a sense of dislike or fear for various things. Blood Insect Trypophobia Jirai Things to be avoided because of dislike or offence 3
Introduction In comics we are more likely to encounter Jirai than in everyday life. Insects Murder Grotesque 4
Questionnaire survey on Jirai Are there any expressions you dislike in comics? (876 respondents) Yes 53.4% No 46.6% grotesque, cruel, sexual, blood, insects, ‥‥ 5
Introduction Jirai are highly reader-dependent. It is not appropriate to regulate the depiction. 6
Introduction People can avoid reading comics if the subject matter can be inferred. ©︎Masaya Hokazono, Yu Satomi 2015 ©︎Hajime Isayama 2010 If Jirai is not the main theme of the comic but only a part of it, it is not easy to notice its existence before reading. 7
Introduction ©︎Masaki Kasahara 2014 Readers can enjoy comics as long as they overlook parts that involve jirai. 8
Purpose To create a system that enables people to read comics without worrying about encountering Jirai. 9
Related work ・Less likely to induce anxiety if aversive stimuli are predictable. [Schmitz et. al 2012] ・Presentation of risk information increases acceptability. [Ono et. al, 2019] By presenting users with information about Jirai in comics beforehand, readers will be able to avoid Jirai. ・Ono, Kyoko, Kato, Etsuko, and Tsunemi, Kiyotaka. (2019) “Does risk information change the acceptance of hydrogen refueling stations in the general Japanese population?” International Journal of Hydrogen Energy 44(31): 16038-16047. ・Schmitz, Anja, and Grillon, Christian. (2012) “Assessing fear and anxiety in humans using the threat of predictable and unpredictable aversive events (the NPU-threat test).” Nat Protoc 7(3): 527-532. 10
Method A system that allows users to flag Jirai while reading and share the information to announce its appearance and location. Flag: Bugs Database Bugs will appear on the next page. Readers who dislike bugs Flag: Bugs Blood will appear on the next page. Flag: Blood Readers who dislike blood 11
Implementation : Flagging Click on the screen opens a window to flag. ©︎Taiga Sassa 2015 12
Implementation : Flagging Select the relevant Jirai and its position. ©︎Taiga Sassa 2015 13
Implementation : Announcement If the user is about to encounter a registered Jirai, a warning will be displayed. ©︎Taiga Sassa 2015
Exp1. Jirai flagging experiments Investigating the ease of Jirai flagging and individual differences. Participants : 11 university students Subjects Jirai : insects Jirai flagging Questionnaire survey Assigning Jirai flags while reading comics using the system. 15
Exp1. Jirai flagging experiments Investigating the ease of Jirai flagging and individual differences. Participants : 11 university students Subjects Jirai : insects Jirai flagging Questionnaire survey Investigated whether they dislike insects and the ease of use of the system. 16
Exp1. Jirai flagging experiments｜Result Ease of use rating averaged 4.09 on a 5-point scale. Comic A Comic B Description perceived as Jirai varies among participants. 17
Exp2. Long-term operational experiments | Overview Verified the system’s practicality and impact on the reading experience. Participants : 17 university students Period : 4 weeks System operation Questionnaire survey Using the system to read any chosen comic for four weeks. 18
Exp2. Long-term operational experiments | Results Participants : 17 university students Period : 4 weeks Comics read: 372 Flags added: 393 Announcement made: 105 • Through the experiment, we were able to operate the system as a Jirai collection and announcement service. • The number of announcements varied among participants. We could collect reader-dependent Jirai in the system. 19
Exp2. Long-term operational experiments | Results Positive feedback from the survey I thought it was good that there was a warning so that I would not be surprised when I opened the page. I was prepared for scary depictions because of the genre, but they were more unpleasant than I expected, so the warning helped me prepare myself for reading the book. 20
Exp2. Long-term operational experiments | Results Negative feedback from the survey I felt a little uncomfortable reading a comic with many Jirai scenes because they appeared in the middle of reading. Multiple warnings of Jirai can interfere with the reader's concentration while reading. 21
Discussion • Even during the four-week operation, the system successfully announced Jirai flagged by other users. • The system's Jirai announcement helped users mentally prepare for the Jirai. • In some cases, the frequent announcement may have interfered with the user’s concentration while reading. 22
Discussion Consideration of Jirai flagging by ChatGPT. There are individual differences in the criteria for flagging. A large number of people are needed to cover a wide range of comics. Automatic flagging using chatGPT (Vision API). 23
Discussion Consideration of Jirai flagging by ChatGPT. We selected two comics containing insects from the Manga109 and manually annotated them. GPT was used to determine whether insects were present in each page. Accuracy Recall Precision Comic 1 72.16% 70.59% 58.54% Comic 2 90.29% 66.67% 18.18% Improving the AI’s recognition accuracy could replace human flagging. 24
Discussion True-Positive ©︎Tatsuki Nouda 1993 Consideration of Jirai flagging by ChatGPT. True-Negative False-Positive False-Positive 25
Discussion Consideration of Jirai flagging by ChatGPT. Load comic images (Manga109) and have GPT evaluate whether they contain insect Jirai. Validate the accuracy using the correct data manually determined by the authors. Accuracy Recall Precision Comic 1 72.16% 70.59% 58.54% Comic 2 90.29% 66.67% 18.18% Improving the AI’s recognition accuracy could replace human flagging. 26
Future work • Adjust the announcement frequency to avoid disrupting the reading experience. • Improve the UI to maintain immersion. • Make the system more user-friendly by making it compatible with smartphone applications. 27
Summary PURPOSE Realizing a way to read comics without worrying about Jirai. METHOD Flagging Jirai and presenting a warning to other readers who have the same Jirai. EXPERIMENT Examine the practicality of the system through long-term experiments. RESULT A certain number of warnings were given, and readers were encouraged to be prepared. 28