RIM_ONE_v3.hdf5 and all_data.hdf5 need how to generate it

[LuShuaie]

RIM_ONE_v3.hdf5 and all_data.hdf5 need how to generate it, what are the requirements?

[seva100]

@2017john, HDF5 datasets can be recreated with scripts/Organize datasets.ipynb notebook.

Nevertheless, I have uploaded all the datasets here.

[LuShuaie]

@seva100 Thank you.

[Geeks-Sid]

Hey @seva100 I am facing a similar issue in this particular line
h5f = h5py.File("../data/hdf5_datasets/all_data.hdf5", "r+")
I cant find this particular dataset at all. not even in the folder you mentioned.

[seva100]

@Geeks-Sid all_data.hdf5 is just a union of all data sets from the folder I mentioned.
For example, if you need DRIONS-DB data, you can just replace it with DRIONS-DB.hdf5

[Geeks-Sid]

@seva100 Hey, Can you explain how to create a union of all the datasets?

[seva100]

@Geeks-Sid you can just copy items from all the datasets by something like this:

import h5py

h5f_in = h5py.File("DRIONS-DB.hdf5", "r")
h5f_out = h5py.File("all_data.hdf5", "w")
h5f_out['DRIONS-DB/orig/images'] = h5f_in['DRIONS-DB/orig/images']
... # copy all other items
h5f_in.close()
h5f_out.close()

Or you can recreate it with scripts/Organize datasets.ipynb notebook.
But I think you won't really need a union of all the datasets for the sake of replication.