Calculate unnormalized Ripley's K function. Notes ----- [1] Center vs Surround ... other_data is CENTER, self._tr

Maybe normalize K values to Complete Spatial Randomness (CSR), center to zero, and/or make them have a unit variance. If we

Check if slidename is not in the index of a dataframe.

Get summary of features from one ROI.

Calculate entropy in collagen fiber length & direction. See: Li H, et al. Collagen fiber orientation disorder from H&E images is

Convert from x, y centroids to xmin, ymin, xmax, ymax.

Run extra feature extraction (eg collagen entropy) for a single slide.

Extract features aware of center (self) and surrounding neighborhood for all distinct tissue region objects.

Extract features aware of center (self) and surrounding neighborhood for a single distinct tissue region object. Parameters

Extract morphology of tissue regions belonging to a class.

Run nuclear feature extraction for a single slide.

Run nuclear feature extraction for a single slide.

Extract morhology of each distinct tissue region in the slide.

Get boundary correction terms for data.

Get cleaned up connected components from a binary mask.

Returns ------- DataFrame Dataframe of features where rows are the roi names and columns are the colu

If certain features are reuqested, construct a K-d tree that allows Ripley's K type of calculation like no of nuclei within a certain

Process the semantic masks for one tile from the semantic WSI mask. The returned list has two elements, the first being the region se

Check if slidename is in the index of a dataframe.

Check if slidename is not in the index of a dataframe.

Summarize nuclei enclosed within region object. Parameters ---------- obj_semantic: np.ndarray semantic segmentat

Summarize the neighborhood of a single object within x pixels. Parameters ---------- objcode: int object

summarize per-fibril edge morphometric features

Summarize the neighborhood of a single tissue region object. Parameters ---------- objcode: int object c

Fit KD tree. Parameters ---------- data : array_like, shape (x, y) The n data points of dimension m to be indexed

Get summary of features from one ROI.

Fast, vectorized IOU components. Modified from: https://medium.com/@venuktan/vectorized-intersection-over-union ... -

DataFrame representation of features (one row per tissue region).

Summarize slide regions as a single row in the dataset.

Fast implementation of Ripley's K using KD-trees. Parameters ---------- radii: Iterable[int] radii at wh

This represents the slide as a feature matrix summarizing the collagen with tumor-associated stroma.

This represents the slide as a feature matrix summarizing its semantic segmantation mask.

This summarizes the slide as a nuclear feature matrix, as well as some basic region pixel counts.

Event density.

Standard deviation of Ripley's K at CSR for various radii at CSR. References: ----------- Lagache T, Lang G, Sau

Scale factor from mask MPP to desired mpp

Labeled masks for each of the region classes with discrete boundaries. Note that we store the arrays to csr_matrix which is a sparse

Perimeter of the region of interest from which events exist.