Skip to content

eyepy.core.filter

filter_by_depth(drusen_map, minimum_depth=2)

Parameters:

Name Type Description Default
drusen_map NDArray[bool_]
required
minimum_depth int
2

Returns:

Source code in src/eyepy/core/filter.py 
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
def filter_by_depth(drusen_map: npt.NDArray[np.bool_],
                    minimum_depth: int = 2) -> npt.NDArray[np.bool_]:
    """

    Args:
        drusen_map:
        minimum_depth:

    Returns:

    """
    filtered_drusen = np.copy(drusen_map)
    if minimum_depth == 0:
        return drusen_map
    # get array where connected components get same label
    connected_component_array, _ = ndimage.label(drusen_map, output=None)
    # Go through each component, sum it along axis 0 and check max depth against threshold
    max_depths = np.zeros_like(connected_component_array)
    for label, drusen_pos in enumerate(
            ndimage.find_objects(connected_component_array)):
        component_sub_vol = connected_component_array[drusen_pos]
        component_max_depth = np.max(
            np.sum(component_sub_vol == label + 1, axis=0))
        component_sub_vol[component_sub_vol == label + 1] = component_max_depth
        max_depths[drusen_pos] = component_sub_vol
    filtered_drusen[max_depths < minimum_depth] = False
    return filtered_drusen.astype(bool)

filter_by_height_enface(drusen_map, minimum_height=2)

Parameters:

Name Type Description Default
drusen_map NDArray[bool_]
required
minimum_height int
2

Returns:

Source code in src/eyepy/core/filter.py 
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
def filter_by_height_enface(drusen_map: npt.NDArray[np.bool_],
                            minimum_height: int = 2) -> npt.NDArray[np.bool_]:
    """

    Args:
        drusen_map:
        minimum_height:

    Returns:

    """
    if minimum_height == 0:
        return drusen_map

    projection = np.sum(drusen_map, axis=1,
                        keepdims=True)  # Shape (n_bscans, width)

    # Find connected components in the enface projection
    connected_component_array, num_drusen = ndimage.label(projection != 0)
    # print(connected_component_array.shape)
    max_heights = np.zeros_like(connected_component_array)
    for drusen_pos in ndimage.find_objects(connected_component_array):
        # Work on subvolume for faster processing
        component_sub_vol = connected_component_array[drusen_pos]
        # Find current label (most frequent label in the subvolume)
        label = np.bincount(component_sub_vol[component_sub_vol != 0]).argmax()
        component_max_height = np.max(
            projection[drusen_pos][component_sub_vol == label])
        # Set drusen region to drusen max height
        max_heights[drusen_pos][component_sub_vol ==
                                label] = component_max_height

    filtered_drusen = np.copy(drusen_map)
    indices = np.nonzero(max_heights < minimum_height)
    filtered_drusen[indices[0], :, indices[2]] = False
    return filtered_drusen.astype(bool)