Mangaka Drawing Style Recommendation System

Thumbnail Image
Date
2017
Authors
Abu Snaina, Ahmed
Mousa, Adli
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Recommendation System, a small feature that plays a major role in the user decision of daily interaction with the internet. Starting from people you may know Facebook recommendation system, to Amazon promotions, recommendation system proved day by day its significant role to the users different segments. Manga are comics created in Japan or by creators in the Japanese language, conforming to a style developed in Japan in the late 19th century.[1] They have a long and complex pre-history in earlier Japanese art.[2] (Wikipedia) Manga is part of the Japanese culture, With a broad range of genres including action-adventure, business and commerce, comedy, detective, historical drama, horror, mystery, romance, science fiction and fantasy, sexuality, sports and games, and suspense, among others[4][5] It got a huge fan base around the world. MyAnimeList represents the biggest fan base community, with a huge collections of mangas sorted in categories. Its current recommendation system depends on two factors, the genre and plot similarity. During the time, along with the uniqueness of the manga culture, Mangakas built their own drawing styles. It is a signature for each manga, may be independent or dependent to the genre and classification the manga falls in. With this huge library of manga and the huge community the Mangakas forms, there is a significant differences in the drawing style from manga to another, even the same mangaka may draw in different styles for the same genre. Fans may - in most cases - like a specific drawing style group among the others. It is difficult to specify the group as set of variables, as it is subjective - depends on the fan opinion. This problem create the need to build a recommendation system that fulfill the fan expectations, to be able to recommend a manga that lands in the same drawing group of a manga the fan already liked. The major issue with the concept of this recommendation system is to define the drawing style itself. Manga is a collection of scenes, some scenes are in boundaries and some scenes within the parent page. Drawing style may include the scene size, how scenes are sorted, the scene drawing features and color gradients. This research will focus on studying the possibility of building a recommendation system based on a set of human observable set of features. The main aim is to create a prediction that is acceptable for the users and suits their interests.
Description
Keywords
Citation