Measurement Module

flyqma.measurement provides tools for quantifying reporter fluorescence intensity within individual cell nuclei.

Image Segmentation

Tools for detecting individual cell nuclei.

class flyqma.measurement.segmentation.Segmentation(image, seed_kws={}, seg_kws={}, exclude_edges=True)

Object for finding nuclear contours within an image.

Seed detection is performed by finding local maxima in a euclidean distance transform of the image foreground mask. Segmentation is achieved via the watershed method.

Attributes:

seeds (np.ndarray[float]) - seeds for segmentation, 2 x N

labels (2D np.ndarray[int]) - segment label mask, number denotes segment ID

segment_ids (1D np.ndarray[int]) - unique segment IDs, length N

cmap (matplotlib.colors.ColorMap) - segment ID colormap, length N+1

static array_to_dict(arr)

Convert array to dictionary.

exclude_edge_segments()

Removes segments overlaying the edge_mask.

exclude_small_segments(min_area=10)

Exclude small segments.

Args:

min_area (float) - minimum contour area

static find_maxima(im, min_distance=1, num_peaks=inf)

Find local maxima of euclidean distance transform.

Args:

im (np.ndarray[bool]) - 2D boolean foreground mask

min_distance (int) - minimum distance separating maxima, px

num_peaks (int) - maximum number of peaks

Returns:

seeds (np.ndarray[float]) - local maxima, shape (N, 2)

static get_borders(im)

Returns boolean array with borders masked as True.

classmethod get_seeds_from_distance(mask, sigma=2, min_distance=1, num_peaks=inf)

Seed detection via euclidean distance transform of binary map.

Args:

mask (nd.ndarray[bool]) - foreground mask

sigma (float) - smoothing applied to euclidean distance mask

min_distance (int) - minimum pixel distance between local maxima

num_peaks (int) - maximum number of local maxima

Returns:

seeds (dict) - {segment_id: (xpos, ypos)} pairs

classmethod get_segment_mask(im, seeds)

Get mask for markers.

Args:

im (np.ndarray[float]) - image to be segmented

seeds (dict) - {segment_id: [x, y]} pairs

property num_objects

Number of unique objects in image.

show(figsize=(15, 15))

Visualize segment label mask.

Args:

figsize (tuple) - figure size

update_cmap()

Use current seeds to build colormap.

watershed(mask, sigma=0.5, watershed_line=True)

Run watershed segmentation to generate segment label mask.

Args:

mask (np.ndarray[bool]) - binary foreground mask

sigma (float) - parameter for smoothing distance mask

watershed_line (bool) - if True, include 1px line between contours

Expression Measurement

Tools for measuring the properties of detected cell nuclei.

class flyqma.measurement.measure.Measurements(im, labels)

Object measures properties of labeled segments within an image.

Attributes:

colordepth (int) - number of color channels

segment_ids (np.ndarray[float]) - ordered segment labels

levels (dict) - {channel: np.ndarray[float]} - expression levels

std (dict) - {channel: np.ndarray[float]} - expression std. deviation

xpos (np.ndarray[float]) - segment centroid x-positions

ypos (np.ndarray[float]) - segment centroid y-positions

voxel_size (np.ndarray[float]) - segment voxel size

build_dataframe()

Build and return dataframe containing all measurements.

Returns:

measurements (pd.DataFrame) - measurement data

static evaluate_centroids(labels)

Evaluate center of mass of each label.

• Note: scipy returns centroids as (y, x) which are flipped to (x, y)

Args:

labels (np.ndarray[int]) - segment label mask

Returns:

center_of_mass (dict) - {segment_id: [xpos, ypos]} pairs

measure_centroids(labels, segment_ids)

Measure the centroid of each segment.

Args:

labels (np.ndarray[int]) - cell segment labels

segment_ids (np.ndarray[int]) - ordered segment IDs

measure_expression(im, labels, segment_ids)

Measure expression levels.

Args:

im (np.ndarray[float]) - 3D array of pixel values

labels (np.ndarray[int]) - cell segment labels

segment_ids (np.ndarray[int]) - ordered segment IDs

measure_segment_size(labels, segment_ids)

Measure the voxel size of each segment.

Args: labels (np.ndarray[int]) - cell segment labels segment_ids (np.ndarray[int]) - ordered segment IDs