Does VIA file need to be genrate manually for each train image

[naresh-nishad]

I am try to mask a circle in my own data set so i need to create a json file using only one image if not i have to do it manually or there is some dirty hack to do this to create json file using whole training dataset