fabioperez/skin-data-augmentation

Source code for the paper 'Data Augmentation for Skin Lesion Analysis' — 🏆 Best Paper Award at the ISIC Skin Image Analysis Workshop @ MICCAI 2018

skin-data-augmentation's Issues

TypeError: join() argument must be str or bytes, not 'int64'

Thanks for your code! But I got a confusing problem: "TypeError: join() argument must be str or bytes, not 'int64".

>  warnings.warn(incorrect_binary_warn % (d, name, 9000, CUDA_VERSION))
Found 33 images from 1 classes.
    Class '1' (0): 33 images.
Found 125 images from 2 classes.
    Class '0' (0): 44 images.
    Class '1' (1): 81 images.
Found 128 images from 2 classes.
    Class '0' (0): 34 images.
    Class '1' (1): 94 images.
Found 128 images from 2 classes.
    Class '0' (0): 34 images.
    Class '1' (1): 94 images.
Found 128 images from 2 classes.
    Class '0' (0): 34 images.
    Class '1' (1): 94 images.
Found 128 images from 2 classes.
    Class '0' (0): 34 images.
    Class '1' (1): 94 images.
ERROR - resnet152-limit-125-full-basic - Failed after 0:04:32!
Traceback (most recent calls WITHOUT Sacred internals):
  File "train.py", line 205, in main
    save_images(datasets['samples'], to=AUGMENTED_IMAGES_DIR, n=32)
  File "train.py", line 116, in save_images
    save_image(dataset[i][0], img_path)
  File "/home/move_eric/auglib-code/skin-data-augmentation-master/auglib/dataset_loader.py", line 48, in __getitem__
    self.data.loc[index, self.image_field])
  File "/home/move_eric/anaconda3/envs/torch2/lib/python3.6/posixpath.py", line 94, in join
    genericpath._check_arg_types('join', a, *p)
  File "/home/move_eric/anaconda3/envs/torch2/lib/python3.6/genericpath.py", line 149, in _check_arg_types
    (funcname, s.__class__.__name__)) from None
TypeError: join() argument must be str or bytes, not 'int64'

I'd really appreciate it if you could help me with this problem! Thank you!

What does the metric 'avp' mean?

As the title says, could you tell me the meaning of it? Thank you very much!

val {'loss': 0.2655325831638038, 'auc': 0.9655381636548296, 'acc': 0.8995361328125, 'avp': 0.9683516626154384}

Suggest to loosen the dependency on sacred

Hi, your project skin-data-augmentation(commit id: 91c6109) requires "sacred==0.7.4" in its dependency. After analyzing the source code, we found that the following versions of sacred can also be suitable, i.e., sacred 0.7.3, since all functions that you directly (3 APIs: sacred.experiment.Experiment.init, sacred.observers.file_storage.FileStorageObserver.create, sacred.observers.telegram_obs.TelegramObserver.from_config) or indirectly (propagate to 22 sacred's internal APIs and 21 outsider APIs) used from the package have not been changed in these versions, thus not affecting your usage.

Therefore, we believe that it is quite safe to loose your dependency on sacred from "sacred==0.7.4" to "sacred>=0.7.3,<=0.7.4". This will improve the applicability of skin-data-augmentation and reduce the possibility of any further dependency conflict with other projects.

May I pull a request to further loosen the dependency on sacred?

By the way, could you please tell us whether such an automatic tool for dependency analysis may be potentially helpful for maintaining dependencies easier during your development?

Final weight for ISIC

Could you kindly release the final model weights for ISIC classification that you mentioned

auglib cannot be found

!pip install pretrainedmodels

!pip freeze

!pip install sacred

!pip install auglib[all]

Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/
ERROR: Could not find a version that satisfies the requirement auglib[all] (from versions: none)
ERROR: No matching distribution found for auglib[all]

muiti-class task

Thanks for your code，I want to use your code to do my own task, which is multi-task classification problem, so could it be useful for multiclass task? Thank you very much if you answer my question!

IndexError: index -1 is out of bounds for axis 0 with size 0

Hi, @fabioperez

I met this error when run generate_lesions.py

Kerb:skin-data-augmentation kerbal$ python3 generate_lesions.py $TRAIN_CSV $TRAIN_DIR 42000 --sigma 20 --output mix_lesion_images
ISIC_0011205 ISIC_0014639
ISIC_0000468 ISIC_0013360
ISIC_0000007 ISIC_0000206
ISIC_0000475 ISIC_0011199
ISIC_0000103 ISIC_0009504
Traceback (most recent call last):
File "generate_lesions.py", line 77, in
args.sigma)
File "generate_lesions.py", line 67, in main
gauss_sigma=sigma)
File "/Users/kerbal/skin-data-augmentation/auglib/augmentation/two_lesions.py", line 67, in mix_lesions
lesion_bg_masked)
File "/Users/kerbal/skin-data-augmentation/auglib/augmentation/two_lesions.py", line 51, in _histogram_matching
t_quantiles /= t_quantiles[-1]
IndexError: index -1 is out of bounds for axis 0 with size 0

What's wrong to me ??

Thanks in advance.
Best,

from @bemoregt.

what shouldi change if my groundtruth is images not cvs files

ConnectionRefusedError: [Errno 111] Connection refused