imgutils.validate.portrait

Overview:

A model for classifying anime portrait images into 3 classes (person, halfbody, head).

The following are sample images for testing.

This is an overall benchmark of all the portrait classification models:

../../_images/portrait_benchmark.plot.py.svg

The models are hosted on huggingface - deepghs/anime_portrait.

anime_portrait_score

Get the scores for different types in an anime portrait.

Parameters:

image (ImageTyping) – The input image.
model_name (str) – The model name. Default is ‘mobilenetv3_v0_dist’.

Returns:

A dictionary with type scores.

Return type:

Dict[str, float]

Examples::

>>> from imgutils.validate import anime_portrait_score
>>>
>>> anime_portrait_score('person/1.jpg')
{'person': 0.999900221824646, 'halfbody': 8.645313209854066e-05, 'head': 1.3387104445428122e-05}
>>> anime_portrait_score('person/2.jpg')
{'person': 0.9999704360961914, 'halfbody': 2.4465465685352683e-05, 'head': 5.071506166132167e-06}
>>> anime_portrait_score('person/3.jpg')
{'person': 0.9999785423278809, 'halfbody': 1.512719154561637e-05, 'head': 6.292278612818336e-06}
>>> anime_portrait_score('halfbody/4.jpg')
{'person': 4.919455750496127e-05, 'halfbody': 0.9999444484710693, 'head': 6.3647335082350764e-06}
>>> anime_portrait_score('halfbody/5.jpg')
{'person': 1.0555699191172607e-05, 'halfbody': 0.9999880790710449, 'head': 1.3210242286731955e-06}
>>> anime_portrait_score('halfbody/6.jpg')
{'person': 1.7451418898417614e-05, 'halfbody': 0.9999822378158569, 'head': 3.2084267331811134e-07}
>>> anime_portrait_score('head/7.jpg')
{'person': 2.7460413321023225e-07, 'halfbody': 1.1532473820352607e-07, 'head': 0.9999996423721313}
>>> anime_portrait_score('head/8.jpg')
{'person': 1.0316136922483565e-07, 'halfbody': 5.840229633236049e-08, 'head': 0.9999998807907104}
>>> anime_portrait_score('head/9.jpg')
{'person': 5.736660568800289e-07, 'halfbody': 7.199210472208506e-08, 'head': 0.9999992847442627}

anime_portrait

Get the primary anime portrait type and its score.

Parameters:

image (ImageTyping) – The input image.
model_name (str) – The model name. Default is ‘mobilenetv3_v0_dist’.

Returns:

A tuple with the primary type and its score.

Return type:

Tuple[str, float]

Examples::

>>> from imgutils.validate import anime_portrait
>>>
>>> anime_portrait('person/1.jpg')
('person', 0.999900221824646)
>>> anime_portrait('person/2.jpg')
('person', 0.9999704360961914)
>>> anime_portrait('person/3.jpg')
('person', 0.9999785423278809)
>>> anime_portrait('halfbody/4.jpg')
('halfbody', 0.9999444484710693)
>>> anime_portrait('halfbody/5.jpg')
('halfbody', 0.9999880790710449)
>>> anime_portrait('halfbody/6.jpg')
('halfbody', 0.9999822378158569)
>>> anime_portrait('head/7.jpg')
('head', 0.9999996423721313)
>>> anime_portrait('head/8.jpg')
('head', 0.9999998807907104)
>>> anime_portrait('head/9.jpg')
('head', 0.9999992847442627)