aecelaya / mist Goto Github PK

There is a typo with the variable
It should be:
"use_n4_bias_correction": bool(self.args.use_n4_bias_correction)}

Receiving error at the preprocessing phase- UnboundLocalError: local variable 'fg_bbox" referenced before assignment.
How to troubleshoot this issue? I'm not sure if public users are permitted to access/edit preprocess.py script?

When I switch datasets, the preprocessing adds the new dataset to the old dataset directories
numpy/images
numpy/labels

If we can remove these directories to avoid this issue, when running the preprocessing, it would really help.

def create_empty_dir(path):
if not (os.path.exists(path)):
os.makedirs(path)
else:
os.system("rm -rf " + path)
os.makedirs(path)

I trained on one dataset and then did the predictions on another dataset using predict.py. I found that inference with the second dataset, the FOV was changing the model didn't do very well with predicting when that happened. I can show you it visually. Is there a way to make sure the FOV is not changing?

I ran the tensorflow code with the Lits data with the flags

nnunet --pocket --xla -amp --patch-size 128 128 128 epochs 125

and I got the following results

Mean, 0.8993201019273616
Std, 0.1050104076908201
25th Percentile, 0.8935995292553426
Median, 0.9303144316101752
75th Percentile, 0.9463229317287942

How can I improve the mean dice score?

Thanks,

Rachel