remap_classes#

Dataset.remap_classes(class_mapping: dict[int, int], new_names: dict[int, str] | None = None, remove_not_mapped: bool = True, remove_emptied_images: bool = False) → Self[source]#

Remap classes ids and names according to a dictionary

Note

In case of class fusion, the class name of the last category_id with respect to class_mapping order will be deduced.

Note

if remove_not_mapped is True, Classes that are not present in the dictionary are removed from the dataset altogether. Otherwise, they are kept as if the identity mapping was in the bottom of class_mapping for this particular class. For potential class fusion, the name of the unmapped class will be used.

See also

Parameters:

class_mapping – old_id -> new_id mapping
new_names – Optimal new_id -> new_name mapping, essentially the new label_map. If category_id is missing from keys, will deduce it from the former one. Defaults to None.
remove_not_mapped – If set to True, will remove classes that are not in class mapping. Otherwise, keep them as is (with potential class fusion). Defaults to True.
remove_emptied_images – If set to True, will remove from self.images the images that are now empty of annotation. Note that it will keep the images that were empty before the remapping. Defaults to False.

Returns:

New dataset object with updated label maps, category ids and category_names

Example

>>> from lours.utils.doc_utils import dummy_dataset
>>> example = dummy_dataset(2, 2, seed=1)
>>> example
Dataset object containing 2 images and 2 objects
Name :
    shake_effort_many
Images root :
    care/suggest
Images :
    width  height        relative_path  type  split
id
0     955     229  determine/story.jpg  .jpg  train
1     131     840       air/method.bmp  .bmp  train
Annotations :
    image_id category_str  category_id  ...   box_y_min   box_width  box_height
id                                      ...
0          1       listen           14  ...  276.974642    9.718823  184.684056
1          0        reach           22  ...    6.311037  123.141689  174.239136

[2 rows x 8 columns]
Label map :
{14: 'listen', 15: 'marriage', 22: 'reach'}
>>> example.remap_classes({14: 1})
Dataset object containing 2 images and 1 object
Name :
    shake_effort_many
Images root :
    care/suggest
Images :
    width  height        relative_path  type  split
id
0     955     229  determine/story.jpg  .jpg  train
1     131     840       air/method.bmp  .bmp  train
Annotations :
    image_id category_str  category_id  ...   box_y_min  box_width  box_height
id                                      ...
0          1       listen            1  ...  276.974642   9.718823  184.684056

[1 rows x 8 columns]
Label map :
{1: 'listen'}

>>> example.remap_classes({14: 1}, remove_not_mapped=False)
Dataset object containing 2 images and 2 objects
Name :
    shake_effort_many
Images root :
    care/suggest
Images :
    width  height        relative_path  type  split
id
0     955     229  determine/story.jpg  .jpg  train
1     131     840       air/method.bmp  .bmp  train
Annotations :
    image_id category_str  category_id  ...   box_y_min   box_width  box_height
id                                      ...
0          1       listen            1  ...  276.974642    9.718823  184.684056
1          0        reach           22  ...    6.311037  123.141689  174.239136

[2 rows x 8 columns]
Label map :
{1: 'listen', 15: 'marriage', 22: 'reach'}

>>> example.remap_classes(
...     {14: 1},
...     remove_not_mapped=False,
...     new_names={1: "new_listen", 15: "new_marriage"},
... )
Dataset object containing 2 images and 2 objects
Name :
    shake_effort_many
Images root :
    care/suggest
Images :
    width  height        relative_path  type  split
id
0     955     229  determine/story.jpg  .jpg  train
1     131     840       air/method.bmp  .bmp  train
Annotations :
    image_id category_str  category_id  ...   box_y_min   box_width  box_height
id                                      ...
0          1   new_listen            1  ...  276.974642    9.718823  184.684056
1          0        reach           22  ...    6.311037  123.141689  174.239136

[2 rows x 8 columns]
Label map :
{1: 'new_listen', 15: 'new_marriage', 22: 'reach'}

>>> example.remap_classes(
...     {14: 1},
...     remove_emptied_images=True,
... )
Dataset object containing 1 image and 1 object
Name :
    shake_effort_many
Images root :
    care/suggest
Images :
    width  height   relative_path  type  split
id
1     131     840  air/method.bmp  .bmp  train
Annotations :
    image_id category_str  category_id  ...   box_y_min  box_width  box_height
id                                      ...
0          1       listen            1  ...  276.974642   9.718823  184.684056

[1 rows x 8 columns]
Label map :
{1: 'listen'}

Note that only empited images are removed. Images that were already empty before are kept

>>> example = dummy_dataset(2, 2)
>>> example
Dataset object containing 2 images and 2 objects
Name :
    inside_else_memory
Images root :
    such/serious
Images :
    width  height      relative_path   type  split
id
0     342     136       help/me.jpeg  .jpeg  train
1     377     167  whatever/wait.png   .png  train
Annotations :
    image_id category_str  category_id  ...  box_y_min   box_width  box_height
id                                      ...
0          0         step           15  ...  73.932999   71.552480   42.673983
1          0          why           19  ...   4.567638  248.551257  122.602211

[2 rows x 8 columns]
Label map :
{15: 'step', 19: 'why', 25: 'interview'}
>>> example.remap_classes({25: 1}, remove_emptied_images=True)
Dataset object containing 1 image and 0 object
Name :
    inside_else_memory
Images root :
    such/serious
Images :
    width  height      relative_path  type  split
id
1     377     167  whatever/wait.png  .png  train
Annotations :
Empty DataFrame
Columns: [image_id, category_str, category_id, split, box_x_min, box_y_min, box_width, box_height]
Index: []
Label map :
{1: 'interview'}