Dealing with 16 bit images

fastai
1

Hi,
I have 16 bit .tiff images and it seems that these are not loaded correctly. (my version is 1.0.38)

I tried sublcassing ImageItemList with overriding the open method with:

def open_image_16bit2rgb( fn ):
    # step 1 : open 16 bit grayscale and convert to int32 and create a view on the image a np.asarray
    a = np.asarray(PIL.Image.open( fn ).convert('I'))
    #step 2: add an 1 dimension so we have height, width, 1 channel
    a = np.expand_dims(a,axis=2)
    #create two extra channels to make it an rgb image
    a = np.repeat(a, 3, axis=2)
    return Image( pil2tensor(a, np.float32 ).div(65535) )

which I found somewhere else, but it didn’t work.
Thanks

2

Sorry to bother you @sgugger, but I couldn’t find a solution. (I updated to 1.0.42)
So I did this:

class ImageItemList2(ImageItemList):
    def open(self, fn):
        return open_image_16bit2rgb(fn)

I created a databunch and when I do this

db.train_ds[0][0]

everything looks fine. But I get a broken pipe error after I try to load a batch. The dataloader is not created correctly.

  1. How can I fix this?
  2. How can I debug broken pipe errors?

BrokenPipeError Traceback (most recent call last)
in
1 dl = db.train_dl
2 dl2 = dl.dl
----> 3 next(iter(dl2))

\miniconda3\envs\f1\lib\site-packages\torch\utils\data\dataloader.py in iter(self)
817
818 def iter(self):
–> 819 return _DataLoaderIter(self)
820
821 def len(self):

\miniconda3\envs\f1\lib\site-packages\torch\utils\data\dataloader.py in init(self, loader)
558 # before it starts, and del tries to join but will get:
559 # AssertionError: can only join a started process.
–> 560 w.start()
561 self.index_queues.append(index_queue)
562 self.workers.append(w)

\miniconda3\envs\f1\lib\multiprocessing\process.py in start(self)
103 ‘daemonic processes are not allowed to have children’
104 _cleanup()
–> 105 self._popen = self._Popen(self)
106 self._sentinel = self._popen.sentinel
107 # Avoid a refcycle if the target function holds an indirect

\miniconda3\envs\f1\lib\multiprocessing\context.py in _Popen(process_obj)
221 @staticmethod
222 def _Popen(process_obj):
–> 223 return _default_context.get_context().Process._Popen(process_obj)
224
225 class DefaultContext(BaseContext):

\miniconda3\envs\f1\lib\multiprocessing\context.py in _Popen(process_obj)
320 def _Popen(process_obj):
321 from .popen_spawn_win32 import Popen
–> 322 return Popen(process_obj)
323
324 class SpawnContext(BaseContext):

\miniconda3\envs\f1\lib\multiprocessing\popen_spawn_win32.py in init(self, process_obj)
63 try:
64 reduction.dump(prep_data, to_child)
—> 65 reduction.dump(process_obj, to_child)
66 finally:
67 set_spawning_popen(None)

\miniconda3\envs\f1\lib\multiprocessing\reduction.py in dump(obj, file, protocol)
58 def dump(obj, file, protocol=None):
59 ‘’‘Replacement for pickle.dump() using ForkingPickler.’’’
—> 60 ForkingPickler(file, protocol).dump(obj)
61
62 #

BrokenPipeError: [Errno 32] Broken pipe

3

Broken Pipe Errors are often completely unrelated to fastai, and more on the pytoch side, depending on your software. You can get rid of them by setting num_workers=0 in your databunch but it’s going to be slower.

3 Likes
4

Hi! In my case I want to load 16 bits png images for chest x-ray images classification. I’ve checked and Pytorch does normalize in the range [0, 1] all images (8 and 16 bits ones). When calling dls.show_batch() all images looks white, which I think is due to the function is not normalizing correctly the values from the 16 bit images [0, 65535]. I don’t care that much about show_batch() but if the model is getting the images as I want to (16 bits [0, 65535] but normalize in the range [0, 1]).

5

I agree there. OI have a similar problem since the normalization should not be redone by Pytorch. Is there a way to shut it off when the imags are loaded by FastAI? Because until now it is not fitting because the images have too little dynamic range in this way. Possibly the solution is to multiply images by
65000