Box operations
powerboxes.boxes_areas(boxes)
Compute areas of boxes.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
boxes |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
Returns:
| Type | Description |
|---|---|
NDArray[float64]
|
np.ndarray: 1d array of areas |
powerboxes.box_convert(boxes, in_fmt, out_fmt)
Convert boxes from one format to another.
Available formats are
- 'xyxy': (xmin, ymin, xmax, ymax)
- 'xywh': (xmin, ymin, width, height)
- 'cxcywh': (center_x, center_y, width, height)
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
boxes |
NDArray[T]
|
2d array of boxes in in_fmt |
required |
in_fmt |
str
|
format of input boxes |
required |
out_fmt |
str
|
format of output boxes |
required |
Returns:
| Type | Description |
|---|---|
NDArray[T]
|
np.ndarray: boxes in out_fmt |
powerboxes.masks_to_boxes(masks)
Convert masks to boxes in xyxy format.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
masks |
NDArray[bool_]
|
3d array of masks in (N, H, W) format |
required |
Raises:
| Type | Description |
|---|---|
TypeError
|
if masks is not numpy array |
Returns:
| Type | Description |
|---|---|
NDArray[uint64]
|
npt.NDArray[np.uint64]: 2d array of boxes in xyxy format |