COMICS Text+

COMICS Text+ 数据集是由KUIS AI Center, Koç University创建的，旨在提升漫画书中的文本检测和识别技术。该数据集包含两个主要部分：'COMICS Text+: Detection' 和 'COMICS Text+: Recognition'，分别用于训练文本检测和文本识别模型。数据集通过GitHub公开，包含超过200万个文本框的转录，适用于研究包括文本检测、识别以及漫画高级处理等任务。The COMICS Text+ dataset was developed by the KUIS AI Center at Koç University to advance text detection and recognition technologies in comic books. It consists of two core components: "COMICS Text+: Detection" and "COMICS Text+: Recognition", which are respectively used for training text detection and text recognition models. The dataset is publicly hosted on GitHub, contains over 2 million transcribed text bounding boxes, and supports research tasks including text detection, text recognition, and advanced comic book processing.