Skip to content

semsegtargettransformer

semsegtargettransformer

Module for SemSegTargetTransformer

Classes

SemSegTargetTransformer 

SemSegTargetTransformer(default_value=255)

Bases: NetTargetTransformer

transforms classification net targets for Lens Defect SemSeg

Source code in niceml/mlcomponents/targettransformer/semsegtargettransformer.py 
20
21
def __init__(self, default_value: int = 255):
    self.default_value = default_value
Functions
get_single_target 
get_single_target(semseg_data)

returns mask array of one lens

Source code in niceml/mlcomponents/targettransformer/semsegtargettransformer.py 
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
def get_single_target(self, semseg_data: SemSegData) -> np.ndarray:
    """returns mask array of one lens"""
    out_dd: OutputImageDataDescription = check_instance(
        self.data_description, OutputImageDataDescription
    )
    input_dd: InputImageDataDescription = check_instance(
        self.data_description, InputImageDataDescription
    )
    input_image_size: ImageSize = input_dd.get_input_image_size()
    output_image_size: ImageSize = out_dd.get_output_image_size()
    class_count: int = out_dd.get_output_channel_count()
    mask_image = semseg_data.mask_image
    output_array = np.zeros(out_dd.get_output_tensor_shape())
    if out_dd.get_output_image_size().np_array_has_same_size(mask_image):
        output_array[
            mask_image != self.default_value,
            mask_image[mask_image < class_count],
        ] = 1
    else:
        ds_masked_hist = get_downscaled_masked_histogram(
            mask_image=mask_image,
            num_classes=class_count,
            default_value=self.default_value,
            ds_factor=int(input_image_size.get_division_factor(output_image_size)),
        )
        output_array[ds_masked_hist > 0] = 1
    return output_array

Functions