Skip to content

Resize

Resize ¤

Bases: Subfunction_Base

A Resize Subfunction class which resizes an images according to a desired shape.

2D image

Shape have to be defined as tuple with x and y size: Resize(shape=(224, 224))

Resizing is done via albumentations resize transform which uses bi-linear interpolation by default.
https://albumentations.ai/docs/api_reference/augmentations/geometric/resize/

3D volume

Shape has to be defined as tuple with x, y and z size: Resize(shape=(128, 128, 128))

Resizing is done via volumentations resize transform which uses bi-linear interpolation by default.
https://github.com/muellerdo/volumentations

Source code in aucmedi/data_processing/subfunctions/resize.py 
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
class Resize(Subfunction_Base):
    """ A Resize Subfunction class which resizes an images according to a desired shape.

    ???+ info "2D image"
        Shape have to be defined as tuple with x and y size: `Resize(shape=(224, 224))`

        Resizing is done via albumentations resize transform which uses bi-linear interpolation by default. <br>
        https://albumentations.ai/docs/api_reference/augmentations/geometric/resize/

    ???+ info "3D volume"
        Shape has to be defined as tuple with x, y and z size: `Resize(shape=(128, 128, 128))`

        Resizing is done via volumentations resize transform which uses bi-linear interpolation by default. <br>
        https://github.com/muellerdo/volumentations
    """
    #---------------------------------------------#
    #                Initialization               #
    #---------------------------------------------#
    def __init__(self, shape=(224, 224), interpolation=1):
        """ Initialization function for creating a Resize Subfunction which can be passed to a
            [DataGenerator][aucmedi.data_processing.data_generator.DataGenerator].

        Args:
            shape (tuple of int):       Resizing shape consisting of a X and Y size. (optional Z size for Volumes).
            interpolation (int):        Interpolation mode for resizing. Using encoding from cv2:
                                        https://docs.opencv.org/3.4/da/d54/group__imgproc__transform.html
        """
        # Initialize parameter
        params = {"p":1.0, "always_apply":True, "interpolation":interpolation}
        # Select augmentation module and add further parameter depending on dimension
        if len(shape) == 2:
            params["height"] = shape[0]
            params["width"] = shape[1]
            mod = albumentations
        elif len(shape) == 3:
            params["shape"] = shape
            mod = volumentations
        else : raise ValueError("Shape for Resize has to be 2D or 3D!", shape)
        # Initialize resizing transform
        self.aug_transform = mod.Compose([mod.Resize(**params)])
        # Cache shape
        self.shape = shape

    #---------------------------------------------#
    #                Transformation               #
    #---------------------------------------------#
    def transform(self, image):
        # Perform resizing into desired shape
        image_resized = self.aug_transform(image=image)["image"]
        # Return resized image
        return image_resized

__init__(shape=(224, 224), interpolation=1) ¤

Initialization function for creating a Resize Subfunction which can be passed to a DataGenerator.

Parameters:

Name Type Description Default
shape tuple of int

Resizing shape consisting of a X and Y size. (optional Z size for Volumes).

 (224, 224)
interpolation int

Interpolation mode for resizing. Using encoding from cv2: https://docs.opencv.org/3.4/da/d54/group__imgproc__transform.html

 1
Source code in aucmedi/data_processing/subfunctions/resize.py 
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
def __init__(self, shape=(224, 224), interpolation=1):
    """ Initialization function for creating a Resize Subfunction which can be passed to a
        [DataGenerator][aucmedi.data_processing.data_generator.DataGenerator].

    Args:
        shape (tuple of int):       Resizing shape consisting of a X and Y size. (optional Z size for Volumes).
        interpolation (int):        Interpolation mode for resizing. Using encoding from cv2:
                                    https://docs.opencv.org/3.4/da/d54/group__imgproc__transform.html
    """
    # Initialize parameter
    params = {"p":1.0, "always_apply":True, "interpolation":interpolation}
    # Select augmentation module and add further parameter depending on dimension
    if len(shape) == 2:
        params["height"] = shape[0]
        params["width"] = shape[1]
        mod = albumentations
    elif len(shape) == 3:
        params["shape"] = shape
        mod = volumentations
    else : raise ValueError("Shape for Resize has to be 2D or 3D!", shape)
    # Initialize resizing transform
    self.aug_transform = mod.Compose([mod.Resize(**params)])
    # Cache shape
    self.shape = shape