convert_imageset vs convert_annoset

[Wureka JI]

Hi, 

I am new to Caffe.

Following all possible tutorial articles, I have prepared required images and annotation xml files.

Now I am going to generate lmdb. I know I can use convert_imageset or convert_annoset.

My question is, what is the difference between convert_imageset and convert_annoset?

Which one should I use ? or when should I use each other ?

Many thanks.