Skip to content

rgb_intensity

Computes the mean, std, and quantiles of RGB intensity of the labelled objects in img, separately for each channel.

Parameters:

Name Type Description Default
img ndarray

Image to compute properties from. Shape (H, W, 3).

 required
label ndarray

Label image. Shape (H, W).

 None
quantiles Tuple[float, ...]

Quantiles to compute for each object.

 (0.25, 0.5, 0.75)

Returns:

Name Type Description
means Tuple[ndarray, ndarray, ndarray]

Mean intensity of each object for each channel (RGB). Each array shape (N,).
std Tuple[ndarray, ndarray, ndarray]

Standard deviation of each object for each channel (RGB). Each array shape (N,).
quantile_vals Tuple[ndarray, ndarray, ndarray]

Quantile values for each object for each channel (RGB). Each array shape (N, len(quantiles)).

Examples:

>>> from histolytics.data import hgsc_cancer_he, hgsc_cancer_nuclei
>>> from histolytics.utils.raster import gdf2inst
>>> from histolytics.nuc_feats.intensity import rgb_intensity
>>>
>>> he_image = hgsc_cancer_he()
>>> nuclei = hgsc_cancer_nuclei()
>>> neoplastic_nuclei = nuclei[nuclei["class_name"] == "neoplastic"]
>>> inst_mask = gdf2inst(
...     neoplastic_nuclei, width=he_image.shape[1], height=he_image.shape[0]
... )
>>> # Extract RGB intensity features from the neoplastic nuclei
>>> means, stds, quantiles = rgb_intensity(he_image, inst_mask)
>>> # RED channel mean intensity
>>> print(means[0].mean())
    0.3659444588664546
Source code in src/histolytics/nuc_feats/intensity.py 
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
def rgb_intensity(
    img: np.ndarray,
    label: np.ndarray = None,
    quantiles: Tuple[float, ...] = (0.25, 0.5, 0.75),
) -> Tuple[
    Tuple[np.ndarray, np.ndarray, np.ndarray],
    Tuple[np.ndarray, np.ndarray, np.ndarray],
    Tuple[np.ndarray, np.ndarray, np.ndarray],
]:
    """Computes the mean, std, and quantiles of RGB intensity of the labelled objects in
    `img`, separately for each channel.

    Parameters:
        img (np.ndarray):
            Image to compute properties from. Shape (H, W, 3).
        label (np.ndarray):
            Label image. Shape (H, W).
        quantiles (Tuple[float, ...]):
            Quantiles to compute for each object.

    Returns:
        means (Tuple[np.ndarray, np.ndarray, np.ndarray]):
            Mean intensity of each object for each channel (RGB). Each array shape (N,).
        std (Tuple[np.ndarray, np.ndarray, np.ndarray]):
            Standard deviation of each object for each channel (RGB). Each array shape (N,).
        quantile_vals (Tuple[np.ndarray, np.ndarray, np.ndarray]):
            Quantile values for each object for each channel (RGB). Each array shape (N, len(quantiles)).

    Examples:
        >>> from histolytics.data import hgsc_cancer_he, hgsc_cancer_nuclei
        >>> from histolytics.utils.raster import gdf2inst
        >>> from histolytics.nuc_feats.intensity import rgb_intensity
        >>>
        >>> he_image = hgsc_cancer_he()
        >>> nuclei = hgsc_cancer_nuclei()
        >>> neoplastic_nuclei = nuclei[nuclei["class_name"] == "neoplastic"]
        >>> inst_mask = gdf2inst(
        ...     neoplastic_nuclei, width=he_image.shape[1], height=he_image.shape[0]
        ... )
        >>> # Extract RGB intensity features from the neoplastic nuclei
        >>> means, stds, quantiles = rgb_intensity(he_image, inst_mask)
        >>> # RED channel mean intensity
        >>> print(means[0].mean())
            0.3659444588664546
    """
    if label is not None and img.shape[:2] != label.shape:
        raise ValueError(
            f"Shape mismatch: img.shape[:2]={img.shape[:2]}, label.shape={label.shape}"
        )

    p2, p98 = np.percentile(img, (2, 98))
    img = rescale_intensity(img, in_range=(p2, p98), out_range=(0, 1))

    means = [[], [], []]
    std = [[], [], []]
    quantile_vals = [[], [], []]
    for i in np.unique(label)[1:]:
        inst = label == i
        inst_pixels = img[inst]  # shape (num_pixels, 3)
        for c in range(3):
            channel_pixels = inst_pixels[:, c]
            means[c].append(np.mean(channel_pixels))
            std[c].append(np.std(channel_pixels))
            quantile_vals[c].append(np.quantile(channel_pixels, quantiles))

    means = tuple(np.array(m) for m in means)
    std = tuple(np.array(s) for s in std)
    quantile_vals = tuple(np.array(q) for q in quantile_vals)

    return means, std, quantile_vals