Incompatible shape problem with Keras ( segmentation model) for batch_size>1 - keras

Question

I am trying to do semantic segmentation using Unet from segmentation model for multi channel (>3) image.
The code works if the batch_size =1. But if I change the batch_size to other values (e.g. 2) then error occurs (InvalidArgumentError: Incompatible shapes):
---------------------------------------------------------------------------
InvalidArgumentError Traceback (most recent call last)
<ipython-input-19-15dc3666afa8> in <module>
22 validation_steps = 1,
23 callbacks=build_callbacks(),
---> 24 verbose = 1)
25
~/.virtualenvs/sm/lib/python3.6/site-packages/keras/legacy/interfaces.py in wrapper(*args, **kwargs)
89 warnings.warn('Update your `' + object_name +
90 '` call to the Keras 2 API: ' + signature, stacklevel=2)
---> 91 return func(*args, **kwargs)
92 wrapper._original_function = func
93 return wrapper
~/.virtualenvs/sm/lib/python3.6/site-packages/keras/engine/training.py in fit_generator(self, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
1424 use_multiprocessing=use_multiprocessing,
1425 shuffle=shuffle,
-> 1426 initial_epoch=initial_epoch)
1427
1428 #interfaces.legacy_generator_methods_support
~/.virtualenvs/sm/lib/python3.6/site-packages/keras/engine/training_generator.py in fit_generator(model, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
189 outs = model.train_on_batch(x, y,
190 sample_weight=sample_weight,
--> 191 class_weight=class_weight)
192
193 if not isinstance(outs, list):
~/.virtualenvs/sm/lib/python3.6/site-packages/keras/engine/training.py in train_on_batch(self, x, y, sample_weight, class_weight)
1218 ins = x + y + sample_weights
1219 self._make_train_function()
-> 1220 outputs = self.train_function(ins)
1221 if len(outputs) == 1:
1222 return outputs[0]
~/.virtualenvs/sm/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py in __call__(self, inputs)
2659 return self._legacy_call(inputs)
2660
-> 2661 return self._call(inputs)
2662 else:
2663 if py_any(is_tensor(x) for x in inputs):
~/.virtualenvs/sm/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py in _call(self, inputs)
2629 symbol_vals,
2630 session)
-> 2631 fetched = self._callable_fn(*array_vals)
2632 return fetched[:len(self.outputs)]
2633
~/.virtualenvs/sm/lib/python3.6/site-packages/tensorflow_core/python/client/session.py in __call__(self, *args, **kwargs)
1470 ret = tf_session.TF_SessionRunCallable(self._session._session,
1471 self._handle, args,
-> 1472 run_metadata_ptr)
1473 if run_metadata:
1474 proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)
InvalidArgumentError: Incompatible shapes: [2,256,256,1] vs. [2,256,256]
[[{{node loss_1/model_4_loss/mul}}]]
I tried to play around by following different posts in the forum but could not solve it. Here, is a portion of code which runs for batch_size=1.
batch_size = 1 # CHANGING ‘batch_size ‘ value other than 1 gives error
train_image_files = glob(patch_img + "/**/*.tif")
# simple_image_generator() is used to work with multi channel (>3) images (the function is
at the end)
train_image_generator = simple_image_generator(train_image_files,
batch_size=batch_size,
rotation_range=45,
horizontal_flip=True,
vertical_flip=True)
train_mask_files = glob(patch_ann + "/**/*.tif")
train_mask_generator = simple_image_generator(train_mask_files,
batch_size=batch_size)
test_image_files = glob(test_img + "/**/*.tif")
test_image_generator = simple_image_generator(test_image_files,
batch_size=batch_size,
rotation_range=45,
horizontal_flip=True,
vertical_flip=True)
test_mask_files = glob(test_ann + "/**/*.tif")
test_mask_generator = simple_image_generator(test_mask_files,
batch_size=batch_size)
train_generator = (pair for pair in zip(train_image_generator, train_mask_generator))
test_generator = (pair for pair in zip(test_image_generator, test_mask_generator))
.
.
num_channels = 8 # no. of channel
base_model = sm.Unet(backbone_name='resnet34', encoder_weights='imagenet')
inp = Input(shape=( None, None, num_channels))
layer_1 = Conv2D( 3, (1, 1))(inp) # map N channels data to 3 channels
out = base_model(layer_1)
model = Model(inp, out, name=base_model.name)
model.summary()
model.compile(
optimizer = keras.optimizers.Adam(lr=learning_rate),
loss = sm.losses.bce_jaccard_loss,
metrics = ['accuracy',sm.metrics.iou_score]
)
model_history = model.fit_generator(train_generator,
epochs = 1,
steps_per_epoch = 1,
validation_data = test_generator,
validation_steps = 1,
callbacks = build_callbacks(),
verbose = 1)
Additional Information:
I am not using the default imageGenerator provided by keras. I am using ‘simple_image_generator’ (slightly modified)
def simple_image_generator(files, batch_size=32,
rotation_range=0, horizontal_flip=False,
vertical_flip=False):
while True:
# select batch_size number of samples without replacement
batch_files = sample(files, batch_size)
# array for images
batch_X = []
# loop over images of the current batch
for idx, input_path in enumerate(batch_files):
image = np.array(imread(input_path), dtype=float)
# process image
if horizontal_flip:
# randomly flip image up/down
if choice([True, False]):
image = np.flipud(image)
if vertical_flip:
# randomly flip image left/right
if choice([True, False]):
image = np.fliplr(image)
# rotate image by random angle between
# -rotation_range <= angle < rotation_range
if rotation_range is not 0:
angle = np.random.uniform(low=-abs(rotation_range),
high=abs(rotation_range))
image = rotate(image, angle, mode='reflect',
order=1, preserve_range=True)
# put all together
batch_X += [image]
# convert lists to np.array
X = np.array(batch_X)
yield(X)

Answer 1

This error was solved by redefining a new image generator instead of simple_image_generator(). The simple_image_generator() worked well with the shape of the images (8 Bands) but did not cope well with the shape of the mask (1 band ).
During the execution, image_generator had 4 dimensions with [2,256,256,1] ( i.e. batch_size, (image size), bands) BUT mask_generator had 3 dimensions only vs. [2,256,256] (i.e. batch_size,(image size))
So reshaping the mask of [2,256,256] to [2,256,256, 1] solved the issue.