Commit f88644a2 authored by narugo1992's avatar narugo1992
Browse files

dev(narugo): use new model for person

parent 47d39c71

docs/source/api_doc/detect/person_detect.benchmark.py

+8 −6
Original line number Diff line number Diff line
@@ -5,13 +5,14 @@ from imgutils.detect import detect_person 




class PersonDetectBenchmark(BaseBenchmark):

    def __init__(self, level):

    def __init__(self, level, plus):

        BaseBenchmark.__init__(self)

        self.level = level

        self.plus = plus



    def load(self):

        from imgutils.detect.person import _open_person_detect_model

        _ = _open_person_detect_model(level=self.level)

        _ = _open_person_detect_model(level=self.level, plus=self.plus)



    def unload(self):

        from imgutils.detect.person import _open_person_detect_model
@@ -19,15 +20,16 @@ class PersonDetectBenchmark(BaseBenchmark): 


    def run(self):

        image_file = random.choice(self.all_images)

        _ = detect_person(image_file, level=self.level)

        _ = detect_person(image_file, level=self.level, plus=self.plus)





if __name__ == '__main__':

    create_plot_cli(

        [

            ('person (yolov8s)', PersonDetectBenchmark('s')),

            ('person (yolov8m)', PersonDetectBenchmark('m')),

            ('person (yolov8x)', PersonDetectBenchmark('x')),

            ('person plus (yolov8m)', PersonDetectBenchmark('m', True)),

            ('person (yolov8s)', PersonDetectBenchmark('s', False)),

            ('person (yolov8m)', PersonDetectBenchmark('m', False)),

            ('person (yolov8x)', PersonDetectBenchmark('x', False)),

        ],

        title='Benchmark for Anime Person Detections',

        run_times=10,

docs/source/api_doc/detect/person_detect.benchmark.py.svg

deleted100644 → 0
+0 −2460

File deleted.

Preview size limit exceeded, changes collapsed.

docs/source/api_doc/detect/person_detect.demo.py

+5 −2
Original line number Diff line number Diff line
@@ -9,8 +9,11 @@ def _detect(img, **kwargs): 


if __name__ == '__main__':

    image_plot(

        (_detect('genshin_post.jpg'), ''),

        (_detect('nian.png'), 'large scale'),

        (_detect('two_bikini_girls.png'), 'closed faces'),

        (_detect('genshin_post.jpg'), 'multiple'),

        (_detect('mostima_post.jpg'), 'anime style'),

        save_as='person_detect.dat.svg',

        columns=1,

        columns=2,

        figsize=(12, 9),

    )

imgutils/detect/person.py

+6 −4
Original line number Diff line number Diff line
@@ -24,14 +24,14 @@ from ..utils import open_onnx_model 




@lru_cache()

def _open_person_detect_model(level: str = 's'):

def _open_person_detect_model(level: str = 'm', plus: bool = True):

    return open_onnx_model(hf_hub_download(

        'deepghs/imgutils-models',

        f'person_detect/person_detect_best_{level}.onnx'

        f'person_detect/person_detect_{"plus_" if plus else ""}best_{level}.onnx'

    ))





def detect_person(image: ImageTyping, level: str = 's', max_infer_size=1216,

def detect_person(image: ImageTyping, level: str = 'm', plus: bool = True, max_infer_size=1216,

                  conf_threshold: float = 0.3, iou_threshold: float = 0.5):

    """

    Overview:
@@ -41,6 +41,8 @@ def detect_person(image: ImageTyping, level: str = 's', max_infer_size=1216, 
    :param level: The model level being used can be either `s`, `m` or `x`.

        The `s` model runs faster with smaller system overhead, while the `m` model achieves higher accuracy.

        The default value is `s`.

    :param plus: Use plus model. Default is ``True``. This argument is not recommended to use ``False`` unless

        you know what this means.

    :param max_infer_size: The maximum image size used for model inference, if the image size exceeds this limit,

        the image will be resized and used for inference. The default value is `1216` pixels.

    :param conf_threshold: The confidence threshold, only detection results with confidence scores above
@@ -72,5 +74,5 @@ def detect_person(image: ImageTyping, level: str = 's', max_infer_size=1216, 
    new_image, old_size, new_size = _image_preprocess(image, max_infer_size)



    data = rgb_encode(new_image)[None, ...]

    output, = _open_person_detect_model(level).run(['output0'], {'images': data})

    output, = _open_person_detect_model(level, plus).run(['output0'], {'images': data})

    return _data_postprocess(output[0], conf_threshold, iou_threshold, old_size, new_size, ['person'])