How to make VOC dataset with target category and bbox coordinates?

vision
1

At present, there is a dataset. The annotation format is a TXT document with only target category and bbox coordinates. How can I create VOC dataset from these data?