We provide the COmic Onomatopoeia dataset (COO) and the source codes to recognize COO. COO consists of onomatopoeia texts in Japanese comics (Manga109). Please refer to the following Github repository and paper for details.