Box metrics
powerboxes.iou_distance(boxes1, boxes2)
Compute pairwise box iou distances.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
boxes1 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
boxes2 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
Raises:
| Type | Description |
|---|---|
TypeError
|
if boxes1 or boxes2 are not numpy arrays |
ValueError
|
if boxes1 and boxes2 have different dtypes |
Returns:
| Type | Description |
|---|---|
NDArray[float64]
|
np.ndarray: 2d matrix of pairwise distances |
powerboxes.parallel_iou_distance(boxes1, boxes2)
Compute pairwise box iou distances, in parallel.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
boxes1 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
boxes2 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
Raises:
| Type | Description |
|---|---|
TypeError
|
if boxes1 or boxes2 are not numpy arrays |
ValueError
|
if boxes1 and boxes2 have different dtypes |
Returns:
| Type | Description |
|---|---|
NDArray[float64]
|
np.ndarray: 2d matrix of pairwise distances |
powerboxes.giou_distance(boxes1, boxes2)
Compute pairwise box giou distances.
see here for giou distance definition
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
boxes1 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
boxes2 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
Raises:
| Type | Description |
|---|---|
TypeError
|
if boxes1 or boxes2 are not numpy arrays |
ValueError
|
if boxes1 and boxes2 have different dtypes |
Returns:
| Type | Description |
|---|---|
NDArray[float64]
|
np.ndarray: 2d matrix of pairwise distances |
powerboxes.parallel_giou_distance(boxes1, boxes2)
Compute pairwise box giou distances, in parallel.
see here for giou distance definition
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
boxes1 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
boxes2 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
Raises:
| Type | Description |
|---|---|
TypeError
|
if boxes1 or boxes2 are not numpy arrays |
ValueError
|
if boxes1 and boxes2 have different dtypes |
Returns:
| Type | Description |
|---|---|
NDArray[float64]
|
np.ndarray: 2d matrix of pairwise distances |
powerboxes.diou_distance(boxes1, boxes2)
Compute pairwise box diou distances.
DIoU distance is defined in https://arxiv.org/pdf/1911.08287.pdf
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
boxes1 |
NDArray[Union[float32, float64]]
|
2d array of boxes in xyxy format |
required |
boxes2 |
NDArray[Union[float32, float64]]
|
2d array of boxes in xyxy format |
required |
Raises:
| Type | Description |
|---|---|
TypeError
|
if boxes1 or boxes2 are not numpy arrays |
ValueError
|
if boxes1 and boxes2 have different dtypes |
Returns:
| Type | Description |
|---|---|
NDArray[float64]
|
np.ndarray: 2d matrix of pairwise distances |
powerboxes.tiou_distance(boxes1, boxes2)
Compute pairwise box tiou (tracking iou) distances.
see here for tiou definition
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
boxes1 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
boxes2 |
NDArray[T]
|
2d array of boxes in xyxy format |
required |
Raises:
| Type | Description |
|---|---|
TypeError
|
if boxes1 or boxes2 are not numpy arrays |
ValueError
|
if boxes1 and boxes2 have different dtypes |
Returns:
| Type | Description |
|---|---|
NDArray[float64]
|
np.ndarray: 2d matrix of pairwise distances |