search

zhixuhao / unet

unet for image segmentation
MIT License
4.55k stars 2k forks source link

Question about labelling #15

Open romdos opened 6 years ago

romdos commented 6 years ago

Question about training data format. As far as I understood you augment data by using imgwrap tool. But how are the labels changed? How to label augmented training images?

dawnranger commented 6 years ago

The image augmentation codes are in file data.py. I think the docstring of class myAugmentation can answer your question:

A class used to augmentate image
Firstly, read train image and label seperately, and then merge them together for the next process
Secondly, use keras preprocessing to augmentate image
Finally, seperate augmentated image apart into train image and label
romdos commented 6 years ago

For some reasons the TIFF package throws an error: ArgumentError: argument 1: <class 'TypeError'>: wrong type. But I solved it using skimage package. Now it works. Do you know some tools for masks creating? I mean if I have my own data.

eli95 commented 6 years ago

File "D:\Anaconda3\lib\site-packages\libtiff\libtiff_ctypes.py", line 1557, in open tiff = libtiff.TIFFOpen(filename, mode)

ArgumentError: argument 1: <class 'TypeError'>: wrong type

find this error when running the split_merge_tif.py how can i solve it

AndrewTotsky commented 6 years ago

@ELI95 Have you solved your problem?