plot_boxes

plot_boxes(
    images,
    boxes,
    truth=None,
    which=None,
    min_score=BOX_MIN_SCORE,
    colours=None,
    labels=None,
)

Detections drawn on the images they were found in, one row per image.

Parameters

Name	Type	Description	Default
images	np.ndarray	one image or a stack.	required
boxes	list[dict[str, Any]]	what `DetectionEngine.predict()` returned - one record per image, with `boxes`, `labels` and `scores` in that image’s own pixels.	required
truth	list[dict[str, Any]] \| None	the annotation in the same shape, drawn in a second colour.	`None`
which	list[int] \| None	which images to draw; the default is the first four.	`None`
min_score	float	boxes below this are not drawn. Not the engine’s `operating_point`, which sits near zero so average precision can integrate a whole ranking.	`BOX_MIN_SCORE`
colours	dict[str, str] \| None	keyed `prediction` and `truth`; defaults to the engine’s.	`None`
labels	list[str] \| None	a row label each.	`None`

Returns

Name	Type	Description
	Figure	A matplotlib `Figure`.

Raises

Name	Type	Description
	PlatypusError	if there is not one record per image.

import numpy as np found = [{“boxes”: [[1, 1, 5, 5]], “labels”: [“cell”], “scores”: [0.91]}] figure = plot_boxes(np.full((1, 8, 8, 3), 0.5), found) len(figure.axes[0].patches), figure.axes[0].texts[0].get_text() (1, ‘cell 0.91’)