Cursive Video Text Dataset (CVTD)

This dataset is the collection of video frames. We have collected videos frames from four different prominent Pakistani News channels. the dataset CVTD contain more then 2500 original images with their labelled files and other dataset CCVTD contain 4000 cropped image frames with their labelled files.

本数据集为视频帧合集。我们从四家颇具影响力的巴基斯坦新闻频道采集了视频帧数据。其中，CVTD数据集包含超过2500张原始图像及其对应的标注文件；另一数据集CCVTD则包含4000张裁剪后的视频帧图像及其对应的标注文件。