TY - DATA T1 - Ot & Sien, a dataset to help the development of object detection in children's book illustrations PY - 2023/06/23 AU - Haoran Wang AU - Seyran Khademi UR - DO - 10.4121/d1f3ca5c-f1e4-48f5-9a04-0564572d2b9c.v1 KW - children books KW - illustrations KW - object detection KW - recognition KW - machine learning KW - imagery content KW - computer vision N2 - <p>The original dataset is Ot & Sien Dataset (https://lab.kb.nl/dataset/ot-sien-dataset). We corrected mistakes and made it ML-ready.</p><p>The purpose of this dataset is to help the development of automatic visual object detection in children's book illustrations. The properties of our dataset are summarized as: </p><ul><li>The dataset consists of illustrations rather than standard photos. </li><li>1452 images with 8241 objects (5.7 per image) are annotated including the category and bounding boxes.</li><li>All images are resized to 416 x 416 with black fitting edges to adapt to the training procedure.</li><li>The dataset follows a natural long-tail property, with some object categories being rare.</li><li>The dataset has imbalanced categories.</li></ul> ER -