Source code for imgutils.detect.face

"""
Overview:
    Detect human faces in anime images.

    Trained on dataset `Anime Face CreateML <https://universe.roboflow.com/my-workspace-mph8o/anime-face-createml>`_
    with YOLOv8.

    .. image:: face_detect_demo.plot.py.svg
        :align: center

    This is an overall benchmark of all the face detect models:

    .. image:: face_detect_benchmark.plot.py.svg
        :align: center

    The models are hosted on
    `huggingface - deepghs/anime_face_detection <https://huggingface.co/deepghs/anime_face_detection>`_.

"""
from typing import List, Tuple, Optional

from ..data import ImageTyping
from ..generic import yolo_predict

_REPO_ID = 'deepghs/anime_face_detection'


[docs]def detect_faces(image: ImageTyping, level: str = 's', version: str = 'v1.4', model_name: Optional[str] = None, conf_threshold: float = 0.25, iou_threshold: float = 0.7) \ -> List[Tuple[Tuple[int, int, int, int], str, float]]: """ Detect human faces in anime images using YOLOv8 models. This function applies a pre-trained YOLOv8 model to detect faces in the given anime image. It supports different model levels and versions, allowing users to balance between detection speed and accuracy. :param image: The input image for face detection. Can be various image types supported by ImageTyping. :type image: ImageTyping :param level: The model level to use. Can be either 's' (standard) or 'n' (nano). The 'n' model is faster with less system overhead, while 's' offers higher accuracy. Default is 's'. :type level: str :param version: The version of the model to use. Available versions are 'v0', 'v1', 'v1.3', and 'v1.4'. Default is 'v1.4'. :type version: str :param model_name: Optional custom model name. If provided, it overrides the auto-generated model name. :type model_name: Optional[str] :param conf_threshold: The confidence threshold for detections. Only detections with confidence scores above this threshold will be returned. Default is 0.25. :type conf_threshold: float :param iou_threshold: The Intersection over Union (IoU) threshold for non-maximum suppression. Detections with IoU above this threshold will be merged. Default is 0.7. :type iou_threshold: float :return: A list of detected faces. Each face is represented by a tuple containing: - Bounding box coordinates as (x0, y0, x1, y1) - The string 'face' (as this function only detects faces) - The confidence score of the detection :rtype: List[Tuple[Tuple[int, int, int, int], str, float]] Examples:: >>> from imgutils.detect import detect_faces, detection_visualize >>> >>> image = 'mostima_post.jpg' >>> result = detect_faces(image) # detect it >>> result [ ((29, 441, 204, 584), 'face', 0.7874319553375244), ((346, 59, 529, 275), 'face', 0.7510495185852051), ((606, 51, 895, 336), 'face', 0.6986488103866577) ] >>> >>> # visualize it >>> from matplotlib import pyplot as plt >>> plt.imshow(detection_visualize(image, result)) >>> plt.show() """ return yolo_predict( image=image, repo_id=_REPO_ID, model_name=model_name or f'face_detect_{version}_{level}', conf_threshold=conf_threshold, iou_threshold=iou_threshold, )