Skip to content

maskvisualizer

maskvisualizer

Module for MaskVisualizer

Classes

MaskVisualizer 

MaskVisualizer(
    hide_gt=False,
    hide_gt_over_threshold=True,
    iou_threshold=0.5,
    match_gt_pred=True,
    **kwargs
)

Bases: ImageVisualizer

Visualizer for images with SemSegInstanceLabels (prediction, ground truth)

Source code in niceml/dashboard/visualizers/maskvisualizer.py 
18
19
20
21
22
23
24
25
26
27
28
29
30
def __init__(
    self,
    hide_gt: bool = False,
    hide_gt_over_threshold: bool = True,
    iou_threshold: float = 0.5,
    match_gt_pred: bool = True,
    **kwargs,
):
    super().__init__(**kwargs)
    self.match_gt_pred = match_gt_pred
    self.hide_gt = hide_gt
    self.hide_gt_over_threshold = hide_gt_over_threshold
    self.iou_threshold = iou_threshold
Functions
get_images_with_labels 
get_images_with_labels(image_data_container)

Returns images of 'image_data_container' with drawn prediction and ground truth labels

Parameters:

  • image_data_container (ImageContainer) –

    Container with an images label information

Returns:

  • List[ndarray] –

    List of images with prediction and ground truth labels drawn on them

Source code in niceml/dashboard/visualizers/maskvisualizer.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
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
def get_images_with_labels(
    self, image_data_container: ImageContainer
) -> List[np.ndarray]:
    """
    Returns images of 'image_data_container' with drawn prediction and ground truth labels

    Args:
        image_data_container: Container with an images label information

    Returns:
        List of images with prediction and ground truth labels drawn on them
    """

    if self.match_gt_pred:
        (
            image_data_container.predictions,
            image_data_container.ground_truth,
        ) = get_kind_of_label_match(
            pred_label_list=image_data_container.predictions,
            gt_label_list=image_data_container.ground_truth,
            hide_gt_over_thresh=self.hide_gt_over_threshold,
            iou_threshold=self.iou_threshold,
        )

    images = []
    for image_path in image_data_container.get_image_paths():
        try:
            images.append(
                self.image_loader(
                    image_path, image_data_container.image_visu_size
                ).astype(np.uint8)
            )
        except FileNotFoundError:
            image = Image.new(
                mode="L",
                size=image_data_container.image_visu_size.to_pil_size(),
            )
            image = np.array(image, dtype=np.uint8)
            images.append(image)

    images = [Image.fromarray(image).convert("RGB") for image in images]

    scale_factor = image_data_container.image_visu_size.get_division_factor(
        image_data_container.model_output_size
    )

    image_data_container = image_data_container.scale_instance_labels(
        scale_factor=scale_factor
    )

    draw_images = []

    for image in images:
        draw_image = draw_labels_on_image(
            image=image,
            pred_error_mask_label_list=image_data_container.predictions,
            gt_error_mask_label_list=image_data_container.ground_truth,
            hide_gt=self.hide_gt,
        )
        draw_images.append(draw_image)

    return [np.array(draw_image) for draw_image in draw_images]

Functions