MONAI Transform stuck converting 4D NIfTI volumes for 3D multichannel network

[AlexWang05]

I am building off of MONAI's 3D segmentation tutorial to work with 4D NIfTI data, where the fourth dimension represents the channels to be inputted for the proposed 3D network. I have adapted the tutorial to better segment with MONAI's DynUNet(nnUNet), but am facing trouble correctly transforming the data into the desired format to train my 3D network in multichannel.

My current approach seems to cause the previously-working DynUNet to get stuck while loading data (estimated time 12+ hrs to load and was Killed by server, was ~1 min previously). I am unable to find if I am transforming/preparing the data correctly for 3D multichannel training.

The current input dimension looks like [num_px_x, num_px,y, num_slices, num_channels], and I hope to transform it into a 3D volume able to be used for a multichannel network.
If helpful, the 4th dimension is of length 7, where index 0 represents an intensity value and indices 1-6 represent a one-hot encoded sequence.

A snippet of my function to get transform:

def get_xforms(mode="train", keys=("image", "label")):
    """returns a composed transform for train/val/infer."""

    xforms = [
        LoadImaged(keys),
        EnsureChannelFirstd(keys='image'),

        AsChannelFirstd(keys),

        Orientationd(keys, axcodes="LPS"),
    ]
    return monai.transforms.Compose(xforms)

Training data loader

# format: [ {'image': 'ct file path', 'label': 'seg file path'} ]
train_files = [{keys[0]: img, keys[1]: seg} for img, seg in zip(images[:n_train], labels[:n_train])]
val_files = [{keys[0]: img, keys[1]: seg} for img, seg in zip(images[-n_val:], labels[-n_val:])]

keys = ("image", "label")
batch_size = 2

train_transforms = get_xforms("train", keys)
train_ds = monai.data.CacheDataset(data=train_files, transform=train_transforms)
train_loader = monai.data.DataLoader(
    train_ds,
    batch_size=batch_size,
    shuffle=True,
    num_workers=2,
    pin_memory=torch.cuda.is_available(),
)

Network function

def get_net():
    kernels=[[3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3]]
    strides=[[1, 1, 1], [2, 2, 2], [2, 2, 2], [2, 2, 2]]

    num_classes = 2
    net = monai.networks.nets.DynUNet(
        spatial_dims=3,
        in_channels=7,
        out_channels=num_classes,
        kernel_size=kernels,
        strides=strides,
        upsample_kernel_size=strides[1:],
    )

    return net

Other code segments are mostly consistent with MONAI's tutorial

[KumoLiu]

Hi @AlexWang05,
It takes 12+ hrs to load data seems the data is too large to load into memory. I'm not sure if I really understand. You want to transform the data with shape [num_px_x, num_px_y, num_slices, num_channels] to 3D volume, you can just transpose your channel dimension(the last dim) to the first dim, don't need both use EnsureChannelFirstd and AsChannelFirstd. Your output from dataloader should be with shape 7*num_px_x*num_px_y*num_slices. Check your data shape first and maybe reduce your batch size. Hope it can help you!

[AlexWang05]

Hi @KumoLiu and thanks for the reply! I've removed those sections and added TransposeD(keys, (0,3)), in my get_xforms() function. However, I'm still getting ValueError: axes don't match array errors when applying the Transform. I'm unsure of how to check the data dimensions in the current code structure (create data loader with transform as parameter, create and run SupervisedTrainer with train loader as parameter, etc).

I'll keep searching for a fix that aligns with your comment; thanks!

[KumoLiu]

Hi @AlexWang05
You may just check your data like this:

from monai.utils import first
check_data = first(train_loader )
print(check_data["image"].shape, check_data["label"].shape)

Hope it helps you, thanks!

[AlexWang05]

Thanks! The script actually gets stuck before the train_loader definition, at train_ds = monai.data.CacheDataset(data=train_files, transform=train_transforms) with the tqdm output starting with "Loading dataset"

I'll be sure to try to work something out, though! Thanks again!

[KumoLiu]

You may also try to iterate your dataset to check your input data shape. If you still have questions, feel free to ask us!