A study on classification of manga authors and titles based on manga character classification

In recent years, the unauthorized use of manga images has posed a significant issue, causing substantial harm due to the infringement of rights. This study aims to address the unauthorized reproduction of manga, specifically focusing on methods to identify manga pages based on characters appearing within them. Our approach involved recognizing and extracting character face regions as individual images, followed by character identification using machine learning techniques for image classification. These results are significant for identifying manga pages, as they enable the organization of characters’ appearances and their frequency throughout the manga content.