There are no instance masks in dataset bbbc010-2012

[lphilomena]

[lphilomena]

There is only an instance mask in the generated mask folder in ./data/bbbc010-2012/train/masks by the python file "/examples/2d/bbbc010-2012/01-data.ipynb for the bbbc010 dataset, could you please share the masks including all instances in a mask image?

[lmanan]

Hello @lphilomena

So the 01-data.ipynb notebook downloads data from this url https://github.com/juglab/EmbedSeg/releases/download/v0.1.0/bbbc010-2012.zip.

If you manually download from this url and unzip this folder, you would see a directory called train/masks containing 100 files.

My guess is that if you delete the directory bbbc010-2012 within data and try running the notebook again, the problem is solved. Let me know if that is not the case.

[lphilomena]

Thanks for your reply. There is a directory called masks containing 100 files. However, there is only one instance in each file, while there are a few objects in the unsegmented image. Are there mask files with the same number of objects as in the unsegmented images? Many thanks.

Manan Lalit @.***> 于2022年11月29日周二 17:33写道：

Hello @lphilomena https://github.com/lphilomena So the 01-data.ipynb downloads data from this url https://github.com/juglab/EmbedSeg/releases/download/v0.1.0/bbbc010-2012.zip. If you unzip this folder, you would see a directory callled masks containing 100 files. can you check this?

— Reply to this email directly, view it on GitHub https://github.com/juglab/EmbedSeg/issues/24#issuecomment-1330340613, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFUX3FENNIXRGKB2TVMCQHDWKXEX3ANCNFSM6AAAAAARV4YSIM . You are receiving this because you were mentioned.Message ID: @.***>

-- Best Regards! Liu Ping

[lmanan]

Hi @lphilomena These 100 masks are all multi-slice tiffs, with the number of slices in each mask equal to the number of worms present in the corresponding image.

In each individual slice, we can see one of the worms. We could visualize a different worm mask by selecting a different slice.

For example, in this code snippet, we load one of the label masks, and by changing the slice parameter we can visualize a different worm mask:

import tifffile
import matplotlib.pyplot as plt
mask_filename = '' # path to one of the label masks 
mask = tifffile.imread(mask_filename)
slice = 1 # change `slice` for a different worm
plt.imshow(mask[slice])

Let me know if there are questions. Thanks.