No semantic type for probability maps

[qin-yu]

Now we have

class SemanticType(Enum):
    """
    Enum class for image types.

    Attributes:
        RAW (str): Reserved for raw images (e.g. microscopy images)
        SEGMENTATION (str): Reserved for segmentation masks
        PREDICTION (str): Reserved for model predictions
        LABEL (str): Reserved for ground truth labels
    """

    RAW = "raw"
    SEGMENTATION = "segmentation"
    PREDICTION = "prediction"
    LABEL = "label"


class ImageType(Enum):
    """
    Enum class for image types.

    Attributes:
        IMAGE (str): Image data
        LABEL (str): Label data
    """

    IMAGE = "image"
    LABEL = "label"

But a prediction from a neural network may not be a probability map, while some functions needs probability maps. The checking now is using:

    if image.semantic_type != SemanticType.PREDICTION:
        gui_logger.warning(
            "The input image is not a boundary probability map. The task will still attempt to run, but the results may not be as expected."
        )

which doesn't guarantee each pixel in image is between 0 and 1. A normalize_01() is usually called somewhere in the code.