Histogram of Exclamation Marks and Its Application for Comics Analysis

摘要

This paper proposes a histogram formed by counting the number of exclamation marks in comic books for comics analysis. Exclamation marks in comic books are used for expressing character's emotion and are depicted in frequently excited scenes. Also they are easy to detect from different comic books written by various authors. Hence we represent each comic book by its distribution of exclamation marks as a histogram and use such histogram for topic change detection and visualization of relationships between comic books. Experimental results on real comic books show that our bold approach has potential for approximating of contents of comic books by using the proposed histograms.