How to create a ground truth data for images dataset with multiple variables.

I have image dataset that contains an object. I want to create a ground truth database which contains the information of the object in that such as its widht, length, orientation, color, area etc. So that later i will use the image along with ground truth to train on CNN network.

Answer 1

[Rishik Ramena]

As far as I understand, you wish to label images for training a CNN. You can do so by using the Image Labeler app. In addition to manually labelling the objects in your image, you can also add sublabels for those objects and export them as ground truth.

[Raza Ali]

Thank you for sharing the links, they are very helpfull. But how can I add length or width of an object in this?

[Rishik Ramena]

You can add them as object attributes.