cv2ext.bboxes package

Module contents

Subpackage containing tools for working with simple bounding boxes.

Functions

bounding()

Get a bounding box which encloses all the given bounding boxes.

calculate_metrics()

Compute accuracy metrics between two sets of bounding boxes.

constrain()

Constrain a bounding box to be within the bounds of an image.

draw_bboxes()

Draw bounding boxes on an image.

euclidean()

Compute the euclidean distance between two bounding boxes.

filter_bboxes_by_region()

Filter a sequence of bounding boxes by a region to contain them in.

iou()

Calculate the intersection over union of two bounding boxes.

ious()

Calculate the intersection over union of a set of bounding boxes.

manhattan()

Compute the manhattan distance between two bounding boxes.

match()

Greedily find matches between two lists of bounding boxes.

mean_ap()

Calculate the mean average precision of a set of bounding boxes.

nms()

Perform non-maximum suppression on a set of bounding boxes.

resize()

Resize a bounding box based on one image size to another.

resize_many()

Resize a set of bounding boxes based on one image size to another.

score_bbox()

Score a bounding box relative to a target bbox.

score_bboxes()

Score a set of bounding boxes relative to a target bbox.

valid()

Check if a bounding box is valid.

within()

Check if a bounding box is within the bounds of an image.

xyxy_to_nxyxy()

Convert bounding boxes from (x1, y1, x2, y2) to normalized (x1, y1, x2, y2).

xyxy_to_xywh()

Convert bounding boxes from (x1, y1, x2, y2) to (x, y, w, h).

xyxy_to_nxywh()

Convert bounding boxes from ;’x1, y1, x2, y2’ to normalized ‘(x, y, w, h)’.

xyxy_to_yolo()

Convert bounding boxes from (x1, y1, x2, y2) to (cx, cy, w, h) in YOLO format.

xywh_to_xyxy()

Convert bounding boxes from (x, y, w, h) to (x1, y1, x2, y2).

xywh_to_nxyxy()

Convert bounding boxes from (x, y, w, h) to normalized (x1, y1, x2, y2).

xywh_to_nxywh()

Convert bounding boxes from (x, y, w, h) to normalized (x, y, w, h).

xywh_to_yolo()

Convert bounding boxes from (x, y, w, h) to (cx, cy, w, h) in YOLO format.

nxyxy_to_xyxy()

Convert bounding boxes from normalized (x1, y1, x2, y2) to (x1, y1, x2, y2).

nxywh_to_xywh()

Convert bounding boxes from normalized (x, y, w, h) to (x, y, w, h).

nxywh_to_nxywh()

Convert bounding boxes from normalized (x, y, w, h) to normalized (x, y, w, h).

nxyxy_to_yolo()

Convert bounding boxes from normalized (x1, y1, x2, y2) to (cx, cy, w, h) in YOLO format.

nxywh_to_xyxy()

Convert bounding boxes from normalized (x, y, w, h) to (x1, y1, x2, y2).

nxywh_to_nxyxy()

Convert bounding boxes from normalized (x, y, w, h) to normalized (x1, y1, x2, y2).

nxywh_to_xywh()

Convert bounding boxes from normalized (x, y, w, h) to (x, y, w, h).

nxywh_to_yolo()

Convert bounding boxes from normalized (x, y, w, h) to (cx, cy, w, h) in YOLO format.

yolo_to_xyxy()

Convert bounding boxes from YOLO format (cx, cy, w, h) to (x1, y1, x2, y2).

yolo_to_nxyxy()

Convert bounding boxes from YOLO format (cx, cy, w, h) to normalized (x1, y1, x2, y2).

yolo_to_xywh()

Convert bounding boxes from YOLO format (cx, cy, w, h) to (x, y, w, h).

yolo_to_nxywh()

Convert bounding boxes from YOLO format (cx, cy, w, h) to normalized (x, y, w, h).

cv2ext.bboxes.bounding(bboxes: Sequence[tuple[int, int, int, int]]) → tuple[int, int, int, int]

Get a bounding box which encloses all the given bounding boxes.

Parameters:

bboxes (Sequence[tuple[int, int, int, int]]) – A sequence of bounding boxes. Bounding boxes are in form xyxy.

Returns:

The bounding box which encloses all the given bounding boxes.

Return type:

tuple[int, int, int, int]

Raises:

ValueError – If the length of bboxes is zero.

cv2ext.bboxes.calculate_metrics(bboxes1: Sequence[tuple[int, int, int, int] | tuple[tuple[int, int, int, int], float, int]], bboxes2: Sequence[tuple[int, int, int, int] | tuple[tuple[int, int, int, int], float, int]], iou_threshold: float = 0.5, epsilon: float = 1e-06, *, class_agnostic: bool = False) → tuple[list[tuple[int, int]], dict[str, float]]

Compute accuracy metrics between two Sequences of bounding boxes/detections.

Bounding boxes are matched using the greedy algolrithm from match().

Parameters:

• bboxes1 (Sequence[tuple[int, int, int, int] | tuple[tuple[int, int, int, int], float, int]]) – The first Sequence of bounding boxes

• bboxes2 (Sequence[tuple[int, int, int, int] | tuple[tuple[int, int, int, int], float, int]]) – The second Sequence of bounding boxes

• iou_threshold (float, optional) – The IOU threshold which determines whether two bounding boxes are a match. By default, 0.5

• epsilon (float, optional) – The minimum/default value to prevent divide by zero errors. By default, 1e-6

• class_agnostic (bool, optional) – Whether or not to compare class ID (if present) By default, False

Returns:

• tuple[ – list[tuple[int, int]], dict[str, float],

• ] – The list of matching indices and a dict of the computer metrics Metrics are: tp, fp, fn, precision, recall, f1

cv2ext.bboxes.constrain(bbox: tuple[int, int, int, int], image_size: tuple[int, int]) → tuple[int, int, int, int]

Constrain a bounding box to the dimensions of an image.

Coordinates are not assumed to be positive, and the bounding box is not assumed to be within the image at all. As such, it is possible to return (0, 0, 0, 0) or (width, height, width, height) if the bounding box is completely outside the image.

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to constrain. Format is: (x1, y1, x2, y2)

• image_size (tuple[int, int]) – The dimensions of the image. Format is: (width, height)

Returns:

The constrained bounding box. Format is: (x1, y1, x2, y2)

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.draw_bboxes(image: np.ndarray, bboxes: Sequence[tuple[int, int, int, int]], confidences: Sequence[float] | None = None, classes: Sequence[str | int] | None = None, class_map: dict[int, str] | None = None, color: Color | tuple[int, int, int] = Color.RED, thickness: int = 2, opacity: float | None = None, *, copy: bool | None = None) → np.ndarray

Draw bounding boxes on an image.

Parameters:

• image (np.ndarray) – The image to draw the bounding boxes on.

• bboxes (Sequence[tuple[int, int, int, int]]) – The Sequence of bounding boxes to draw. The bounding boxes should be in form: x1, y1, x2, y2 (top-left, bottom-right) format

• confidences (Sequence[float], optional) – The confidence values for each bounding box. If provided, the confidence values will be drawn on the image.

• classes (Sequence[str | int], optional) – The class labels for each bounding box. If provided, the class labels will be drawn on the image with the bounding boxes. If the label is a string, it will be used as is, otherwise it will be used as an index into the class_map.

• class_map (dict[int, str], optional) – The class map to use for converting class indices to labels.

• color (Color, tuple[int, int, int], optional) – The color to draw the bounding boxes. In BGR format and the default is Color.RED.

• thickness (int, optional) – The thickness of the bounding box lines. Default is 2.

• opacity (float, optional) – The opacity to draw the bounding boxes with. By default None or 100%. Can have a high performance impact.

• copy (bool, optional) – Whether or not to copy the image before drawing. Default is False.

Returns:

The image with the bounding boxes drawn.

Return type:

np.ndarray

cv2ext.bboxes.euclidean(bbox1: tuple[int, int, int, int], bbox2: tuple[int, int, int, int]) → float

Compute the euclidean distance between two bboxes.

The computation is performed between the centers of the bounding boxes.

Parameters:

• bbox1 (tuple[int, int, int, int]) – The first bounding box. Bounding box is in form xyxy.

• bbox2 (tuple[int, int, int, int]) – The second bounding box. Bounding box is in form xyxy.

Returns:

The euclidean distance between the two bounding boxes.

Return type:

float

cv2ext.bboxes.filter_bboxes_by_region(bboxes: Sequence[tuple[int, int, int, int]], region: tuple[int, int, int, int], overlap: float = 0.6) → list[tuple[int, int, int, int]]

Get the bounding boxes contained within a region.

Parameters:

• bboxes (Sequence[tuple[int, int, int, int]]) – The bounding boxes in form (x1, y1, x2, y2).

• region (tuple[int, int, int, int]) – The region by which to filter the bounding boxes.

• overlap (float) – The percentage of a bounding boxes area is inside the region to be included.

Returns:

The bounding boxes in the region.

Return type:

list[tuple[int, int, int, int]]

cv2ext.bboxes.iou(bbox1: tuple[int, int, int, int], bbox2: tuple[int, int, int, int]) → float

Calculate the intersection over union of two bounding boxes.

Parameters:

• bbox1 (tuple[int, int, int, int]) – The first bounding box in the format (x1, y1, x2, y2).

• bbox2 (tuple[int, int, int, int]) – The second bounding box in the format (x1, y1, x2, y2).

Returns:

The intersection over union of the two bounding boxes.

Return type:

float

cv2ext.bboxes.ious(bboxes1: list[tuple[int, int, int, int]], bboxes2: list[tuple[int, int, int, int]]) → list[float]

Calculate the intersection over union of two lists of bounding boxes.

Parameters:

• bboxes1 (list[tuple[int, int, int, int]]) – The first list of bounding boxes in the format (x1, y1, x2, y2).

• bboxes2 (list[tuple[int, int, int, int]]) – The second list of bounding boxes in the format (x1, y1, x2, y2).

Returns:

The intersection over union of the two lists of bounding boxes.

Return type:

list[float]

cv2ext.bboxes.manhattan(bbox1: tuple[int, int, int, int], bbox2: tuple[int, int, int, int]) → float

Compute the manhattan distance between two bboxes.

The computation is performed between the centers of the bounding boxes.

Parameters:

• bbox1 (tuple[int, int, int, int]) – The first bounding box. Bounding box is in form xyxy.

• bbox2 (tuple[int, int, int, int]) – The second bounding box. Bounding box is in form xyxy.

Returns:

The manhattan distance between the two bounding boxes.

Return type:

float

cv2ext.bboxes.match(bboxes1: Sequence[tuple[int, int, int, int] | tuple[tuple[int, int, int, int], float, int]], bboxes2: Sequence[tuple[int, int, int, int] | tuple[tuple[int, int, int, int], float, int]], iou_threshold: float = 0.5, *, class_agnostic: bool = False) → list[tuple[int, int]]

Match bounding boxes using a greedy algorithm.

Parameters:

• bboxes1 (Sequence[tuple[int, int, int, int] | tuple[tuple[int, int, int, int], float, int]]) – The first Sequence of bounding boxes

• bboxes2 (Sequence[tuple[int, int, int, int] | tuple[tuple[int, int, int, int], float, int]]) – The second Sequence of bounding boxes

• iou_threshold (float, optional) – The IOU threshold which determines whether two bounding boxes are a match. By default, 0.5

• class_agnostic (bool, optional) – Whether or not to compare class ID (if present) By default, False

Returns:

A list of the matching indices

Return type:

list[tuple[int, int]]

cv2ext.bboxes.mean_ap(bboxes: list[list[tuple[tuple[int, int, int, int], int, float]]], gt_bboxes: list[list[tuple[tuple[int, int, int, int], int]]], num_classes: int, iou_threshold: float = 0.5) → float

Calculate the mean average precision for a set of bounding boxes.

bboxes and gt_bboxes are lists of lists representing the bounding boxes for each image. Each bounding box is represented as a tuple of the form ((x1, y1, x2, y2), class, confidence).

Parameters:

• bboxes (list[list[tuple[tuple[int, int, int, int], int, float]]]) – A list of lists of bounding boxes, each represented as a tuple of the form ((x1, y1, x2, y2), class, confidence

• gt_bboxes (list[list[tuple[tuple[int, int, int, int], int]]]) – A list of lists of ground truth bounding boxes, each represented as a tuple of the form ((x1, y1, x2, y2), class)

• num_classes (int) – The number of classes in the dataset.

• iou_threshold (float, optional) – The threshold for considering a detection a true positive, by default 0.5

Returns:

The mean average precision of the bounding boxes.

Return type:

float

Raises:

• ValueError – If the length of bboxes and gt_bboxes are not equal.

• ValueError – If the length is zero.

cv2ext.bboxes.nms(bboxes: list[tuple[tuple[int, int, int, int], float, int]], iou_threshold: float = 0.5, *, agnostic: bool | None = None) → list[tuple[tuple[int, int, int, int], float, int]]

Perform non-maximum suppression on a list of bounding boxes.

Parameters:

• bboxes (list[tuple[tuple[int, int, int, int], float, int]]) – A list of bounding boxes, each represented as a tuple of the form ((x1, y1, x2, y2), confidence, class

• iou_threshold (float) – The intersection over union threshold for non-maximum suppression.

• agnostic (bool, optional) – If set to True, then bounding boxes of different classes can be compared. By default None, will only compare same classes.

Returns:

A list of bounding boxes, each represented as a tuple of the form ((x1, y1, x2, y2), confidence, class

Return type:

list[tuple[tuple[int, int, int, int], float, int]]

cv2ext.bboxes.nxywh_to_nxyxy(bbox: tuple[float, float, float, float]) → tuple[float, float, float, float]

Convert a bounding box from normalized (x, y, w, h) format to normalized (x, y, x, y).

Parameters:

• bbox (tuple[float, float, float, float]) – The bounding box to convert in normalized (x, y, w, h) format.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in normalized format.

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.nxywh_to_xywh(bbox: tuple[float, float, float, float], image_width: int, image_height: int) → tuple[int, int, int, int]

Convert a bounding box from normalized (x, y, w, h) format to (x, y, w, h).

Parameters:

• bbox (tuple[float, float, float, float]) – The bounding box to convert in normalized (x, y, w, h) format.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in (x, y, w, h) format.

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.nxywh_to_xyxy(bbox: tuple[float, float, float, float], image_width: int, image_height: int) → tuple[int, int, int, int]

Convert a bounding box from normalized (x, y, w, h) format to (xmin, ymin, xmax, ymax).

Parameters:

• bbox (tuple[float, float, float, float]) – The bounding box to convert in normalized (x, y, w, h) format.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in (xmin, ymin, xmax, ymax) format.

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.nxywh_to_yolo(bbox: tuple[float, float, float, float]) → tuple[float, float, float, float]

Convert a bounding box from normalized (x, y, w, h) to YOLO format (x_center, y_center, width, height).

Parameters:

bbox (tuple[float, float, float, float]) – The bounding box to convert in normalized (x, y, w, h) format.

Returns:

The converted bounding box in YOLO format (x_center, y_center, width, height).

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.nxyxy_to_nxywh(bbox: tuple[float, float, float, float], image_width: int, image_height: int) → tuple[float, float, float, float]

Convert a normalized (x, y, x, y) bbox to normalized (x, y, w, h) format.

Parameters:

• bbox (tuple[float, float, float, float]) – The normalized bounding box to convert. The format is (xmin, ymin, xmax, ymax) where values are normalized between 0 and 1.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in normalized (x, y, w, h) format.

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.nxyxy_to_xywh(bbox: tuple[float, float, float, float], image_width: int, image_height: int) → tuple[int, int, int, int]

Convert a normalized (x, y, x, y) bbox to absolute (x, y, w, h) coordinates.

Parameters:

• bbox (tuple[float, float, float, float]) – The normalized bounding box to convert. The format is (xmin, ymin, xmax, ymax) where values are normalized between 0 and 1.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in (x, y, w, h) format.

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.nxyxy_to_xyxy(bbox: tuple[float, float, float, float], image_width: int, image_height: int) → tuple[int, int, int, int]

Convert a normalized (x, y, x, y) bbox to absolute (x, y, x, y) coordinates.

Parameters:

• bbox (tuple[float, float, float, float]) – The normalized bounding box to convert. The format is (xmin, ymin, xmax, ymax) where values are normalized between 0 and 1.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in (xmin, ymin, xmax, ymax) format.

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.nxyxy_to_yolo(bbox: tuple[int, int, int, int]) → tuple[float, float, float, float]

Convert a bounding box from normalized (xmin, ymin, xmax, ymax) to YOLO format (x_center, y_center, width, height).

Parameters:

bbox (tuple[int, int, int, int]) – The bounding box to convert. Format is normalized (xmin, ymin, xmax, ymax).

Returns:

The converted bounding box in YOLO format (x_center, y_center, width, height).

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.resize(bbox: tuple[int, int, int, int], s1: tuple[int, int], s2: tuple[int, int]) → tuple[int, int, int, int]

Resizes a bounding box based on one image size to another.

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to resize. Bounding box is in form xyxy.

• s1 (tuple[int, int]) – The size of the first image. In form (width, height).

• s2 (tuple[int, int]) – The size of the second image. In form (width, height).

Returns:

The resized bounding box. Bounding box is in form xyxy.

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.resize_many(bboxes: list[tuple[int, int, int, int]], s1: tuple[int, int], s2: tuple[int, int]) → list[tuple[int, int, int, int]]

Resizes a list of bounding boxes based on one image size to another.

Parameters:

• bboxes (list[tuple[int, int, int, int]]) – The bounding boxes to resize. Bounding boxes are in form xyxy.

• s1 (tuple[int, int]) – The size of the first image. In form (width, height).

• s2 (tuple[int, int]) – The size of the second image. In form (width, height).

Returns:

The resized bounding boxes. Bounding boxes are in form xyxy.

Return type:

list[tuple[int, int, int, int]]

cv2ext.bboxes.score_bbox(target_bbox: tuple[int, int, int, int], pred_bbox: tuple[int, int, int, int]) → float

Compute a simple score metric for two bounding boxes.

The score is computed as an aggregate between the euclidean distance and the area difference.

Parameters:

• target_bbox (tuple[int, int, int, int]) – The target bounding box. Bounding box is format (xmin, ymin, xmax, ymax)

• pred_bbox (tuple[int, int, int, int]) – The predicted bounding box. Bounding box is format (xmin, ymin, xmax, ymax)

Returns:

The score between the two bounding boxes.

Return type:

float

cv2ext.bboxes.score_bboxes(target_bbox: tuple[int, int, int, int], pred_bboxs: list[tuple[int, int, int, int]]) → list[float]

Compute a simple score metric for a target bounding box and a list of predicted bounding boxes.

The score is computed as an aggregate between the euclidean distance and the area difference.

Parameters:

• target_bbox (tuple[int, int, int, int]) – The target bounding box. Bounding box is format (xmin, ymin, xmax, ymax)

• pred_bboxs (list[tuple[int, int, int, int]]) – The list of predicted bounding boxes. Bounding box is format (xmin, ymin, xmax, ymax)

Returns:

The scores between the target bounding box and the list of predicted bounding boxes.

Return type:

list[float]

cv2ext.bboxes.valid(bbox: tuple[int, int, int, int], shape: tuple[int, int] | None = None) → bool

Check if a bounding box is valid.

The conditions for a valid bounding box are that: the top-left corner is strictly above the bottom-right corner, and the bounding box has strictly greater than zero area. Can also check if the bounding box is within the bounds of an image.

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to check. Bounding box is in form xyxy.

• shape (tuple[int, int], optional) – The shape of the image. If provided, will check if the bounding box is within the bounds of the image.

Returns:

True if the bounding box is valid, False otherwise.

Return type:

bool

cv2ext.bboxes.within(bbox: tuple[int, int, int, int], shape: tuple[int, int]) → bool

Check if a bounding box is within the bounds of an image.

The conditions for a bounding box to be within the bounds of an image are that the top-left corner is within the bounds of the image, and the bottom-right corner is within the bounds of the image.

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to check. Bounding box is in form xyxy.

• shape (tuple[int, int]) – The shape of the image.

Returns:

True if the bounding box is within the bounds of the image, False otherwise.

Return type:

bool

cv2ext.bboxes.xywh_to_nxywh(bbox: tuple[int, int, int, int], image_width: int, image_height: int) → tuple[float, float, float, float]

Convert a bounding box from (x, y, w, h) to normalized (x, y, w, h) format.

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to convert. Bounding box is format (x, y, w, h), where (x, y) is the top-left corner and (w, h) is the width and height.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in normalized (x, y, w, h) format.

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.xywh_to_nxyxy(bbox: tuple[int, int, int, int], image_width: int, image_height: int) → tuple[float, float, float, float]

Convert a bounding box from (x, y, w, h) to normalized (x, y, x, y).

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to convert. Bounding box is format (x, y, w, h), where (x, y) is the top-left corner and (w, h) is the width and height.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in normalized format.

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.xywh_to_xyxy(bbox: tuple[int, int, int, int]) → tuple[int, int, int, int]

Convert a bounding box from (x, y, w, h) to (xmin, ymin, xmax, ymax).

Parameters:

bbox (tuple[int, int, int, int]) – The bounding box to convert. Bounding box is format (x, y, w, h), where (x, y) is the top-left corner and (w, h) is the width and height.

Returns:

The converted bounding box. Bounding box is format (xmin, ymin, xmax, ymax), where (xmin, ymin) is the top-left corner and (xmax, ymax) is the bottom-right corner.

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.xywh_to_yolo(bbox: tuple[int, int, int, int], image_width: int, image_height: int) → tuple[float, float, float, float]

Convert a bounding box from (x, y, w, h) to YOLO format (x_center, y_center, width, height).

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to convert. Format is (x, y, w, h).

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in YOLO format (x_center, y_center, width, height).

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.xyxy_to_nxywh(bbox: tuple[int, int, int, int], image_width: int, image_height: int) → tuple[float, float, float, float]

Convert a bounding box from (xmin, ymin, xmax, ymax) to normalized (x, y, w, h) format.

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to convert. Bounding box is format (xmin, ymin, xmax, ymax), where (xmin, ymin) is the top-left corner and (xmax, ymax) is the bottom-right corner.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in normalized (x, y, w, h) format.

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.xyxy_to_nxyxy(bbox: tuple[int, int, int, int], image_width: int, image_height: int) → tuple[float, float, float, float]

Convert a bounding box from (x, y, x, y) to normalized (x, y, x, y).

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to convert. Bounding box is format (xmin, ymin, xmax, ymax), where (xmin, ymin) is the top-left corner and (xmax, ymax) is the bottom-right corner.

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in normalized format.

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.xyxy_to_xywh(bbox: tuple[int, int, int, int]) → tuple[int, int, int, int]

Convert a bounding box from (xmin, ymin, xmax, ymax) to (x, y, w, h).

Parameters:

bbox (tuple[int, int, int, int]) – The bounding box to convert. Bounding box is format (xmin, ymin, xmax, ymax), where (xmin, ymin) is the top-left corner and (xmax, ymax) is the bottom-right corner.

Returns:

The converted bounding box. Bounding box is format (x, y, w, h), where (x, y) is the top-left corner and (w, h) is the width and height.

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.xyxy_to_yolo(bbox: tuple[int, int, int, int], image_width: int, image_height: int) → tuple[float, float, float, float]

Convert a bounding box from (xmin, ymin, xmax, ymax) to YOLO format (x_center, y_center, width, height).

Parameters:

• bbox (tuple[int, int, int, int]) – The bounding box to convert. Format is (xmin, ymin, xmax, ymax).

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in YOLO format (x_center, y_center, width, height).

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.yolo_to_nxywh(bbox: tuple[float, float, float, float]) → tuple[float, float, float, float]

Convert a YOLO bounding box format to normalized (x, y, w, h).

Parameters:

bbox (tuple[float, float, float, float]) – The bounding box in YOLO format (x_center, y_center, width, height).

Returns:

The converted bounding box in normalized (x, y, w, h) format.

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.yolo_to_nxyxy(bbox: tuple[float, float, float, float]) → tuple[float, float, float, float]

Convert a YOLO bounding box format to normalized (xmin, ymin, xmax, ymax).

Parameters:

bbox (tuple[float, float, float, float]) – The bounding box in YOLO format (x_center, y_center, width, height).

Returns:

The converted bounding box in normalized (xmin, ymin, xmax, ymax) format.

Return type:

tuple[float, float, float, float]

cv2ext.bboxes.yolo_to_xywh(bbox: tuple[float, float, float, float], image_width: int, image_height: int) → tuple[int, int, int, int]

Convert a YOLO bounding box format to (x, y, w, h).

Parameters:

• bbox (tuple[float, float, float, float]) – The bounding box in YOLO format (x_center, y_center, width, height).

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in (x, y, w, h) format.

Return type:

tuple[int, int, int, int]

cv2ext.bboxes.yolo_to_xyxy(bbox: tuple[float, float, float, float], image_width: int, image_height: int) → tuple[int, int, int, int]

Convert a YOLO bounding box format to (xmin, ymin, xmax, ymax).

Parameters:

• bbox (tuple[float, float, float, float]) – The bounding box in YOLO format (x_center, y_center, width, height).

• image_width (int) – The width of the image.

• image_height (int) – The height of the image.

Returns:

The converted bounding box in (xmin, ymin, xmax, ymax) format.

Return type:

tuple[int, int, int, int]