Handwritten Javanese Characters Sewaka Detection (HJCS_DETC)

Handwritten Javanese Characters Sewaka Detection (HJCS_DETC) Dataset is a dataset of a Javanese character taken from Serat Sewaka Manuscript downloaded from khastara.perpusnas.go.id. This dataset is aimed at detection tasks with YOLO format. This manuscript was chosen because of the clearness of the image in each character which can make the detection better.

There are 60 images of manuscript containing annotated 74 classes of Javanese characters. This dataset is highly densed and imbalanced due to the natural condition of the manuscripts.

爪哇手写字符Sewaka检测数据集（Handwritten Javanese Characters Sewaka Detection, HJCS_DETC）的样本均取自从khastara.perpusnas.go.id平台下载的Serat Sewaka手稿中的爪哇手写字符。本数据集采用YOLO格式标注，专为目标检测任务设计。选取该手稿作为数据源，是因其单字符图像清晰度优异，可有效提升检测模型的性能表现。
该数据集共包含60幅手稿图像，其中标注了74类爪哇字符。由于手稿本身的原始馆藏特性，该数据集存在字符分布密度偏高且类别不均衡的问题。