written by jonggeun park

As a dataset for object detection training, we need 1) images and 2) files containing information about the objects in the images.

An annotation (coordinates, what objects are, etc.) that expresses object information within an image can be expressed in two formats: json or yaml

When you download dataset, the dataset may already contain an annotation as json or yaml, but if the provided annotation is a different format, such as csv, xml, or if you create the dataset yourself, you need to convert it to json or yaml format.

We will change the csv file to json format.