[Python] I got a problem with my U-Net model i don't know how to fix it

I'm building a U-Net model to detect breast cancer, I got the dataset from here: https://www.kaggle.com/datasets/aryashah2k/breast-ultrasound-images-dataset

Even though, all images are of format png, when trying to training my model an error occurs that my images are not of adequate format. here is the error:

---------------------------------------------------------------------------
InvalidArgumentError Traceback (most recent call last)
Cell In[51], line 7
5 train_dataset = image_ds.cache().shuffle(BUFFER_SIZE).batch(BATCH_SIZE)
6 print(image_ds.element_spec)
----> 7 model_history = unet.fit(train_dataset, epochs=EPOCHS)

File ~\anaconda3\Lib\site-packages\keras\src\utils\traceback_utils.py:122, in filter_traceback.<locals>.error_handler(*args, **kwargs)
119 filtered_tb = _process_traceback_frames(e.__traceback__)
120 # To get the full stack trace, call:
121 # `keras.config.disable_traceback_filtering()`
--> 122 raise e.with_traceback(filtered_tb) from None
123 finally:
124 del filtered_tb

File ~\anaconda3\Lib\site-packages\tensorflow\python\eager\execute.py:53, in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name)
51 try:
52 ctx.ensure_initialized()
---> 53 tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
54 inputs, attrs, num_outputs)
55 except core._NotOkStatusException as e:
56 if name is not None:

InvalidArgumentError: Graph execution error:

Detected at node decode_image/DecodeImage defined at (most recent call last):
<stack traces unavailable>
Error in user-defined function passed to MapDataset:3 transformation with iterator: Iterator::Root:refetch::BatchV2::Shuffle::MemoryCacheImpl::Filter:arallelMapV2: Unknown image file format. One of JPEG, PNG, GIF, BMP required.
[[{{node decode_image/DecodeImage}}]]
[[IteratorGetNext]] [Op:__inference_one_step_on_iterator_9520]


The erro only occurs when trying to train the model, and it refers to this function:

def preprocess_image(image, mask, target_size=(256, 256)):
try:
# Decode the image and mask safely
image = tf.io.decode_image(image, channels=3, expand_animations=False)
mask = tf.io.decode_image(mask, channels=1, expand_animations=False)

# Check for undefined or zero dimensions
if image.shape is None or image.shape[0] == 0 or image.shape[1] == 0:
print(f"Error: Image has undefined or zero dimensions: {image.shape}")
return None, None

if mask.shape is None or mask.shape[0] == 0 or mask.shape[1] == 0:
print(f"Error: Mask has undefined or zero dimensions: {mask.shape}")
return None, None

# Ensure the image has exactly 3 channels (RGB)
if image.shape[-1] != 3:
print(f"Error: Image does not have 3 channels (found {image.shape[-1]}).")
return None, None

# Normalize the image to range [0, 1]
image = tf.image.convert_image_dtype(image, tf.float32)

# Resize image and mask to target dimensions (256*256)
image = tf.image.resize(image, target_size, method='nearest')
mask = tf.image.resize(mask, target_size, method='nearest')

# Convert mask to a binary (0 or 1) format for classification tasks
mask = tf.cast(tf.math.reduce_max(mask, axis=-1, keepdims=True) > 0, tf.float32) # Ensure binary mask

return image, mask

except Exception as e:
print(f"Error during preprocessing: {str(e)}")
return None, None

# Apply the preprocessing function to the dataset
image_ds = dataset.map(preprocess_image)

# Filter out None values returned by preprocess_image
image_ds = image_ds.filter(lambda img, mask: img is not None and mask is not None)


I tried multiple ways trying to fix this using chatgpt but nothing seems to work.

Continue reading...