app.py

import os
import cv2
import imghdr
import shutil
import warnings
import numpy as np
import gradio as gr
from dataclasses import dataclass
from mivolo.predictor import Predictor
from utils import is_url, download_file, get_jpg_files, MODEL_DIR

TMP_DIR = "./__pycache__"


@dataclass
class Cfg:
    detector_weights: str
    checkpoint: str
    device: str = "cpu"
    with_persons: bool = True
    disable_faces: bool = False
    draw: bool = True


class ValidImgDetector:
    predictor = None

    def __init__(self):
        detector_path = f"{MODEL_DIR}/yolov8x_person_face.pt"
        age_gender_path = f"{MODEL_DIR}/model_imdb_cross_person_4.22_99.46.pth.tar"
        predictor_cfg = Cfg(detector_path, age_gender_path)
        self.predictor = Predictor(predictor_cfg)

    def _detect(
        self,
        image: np.ndarray,
        score_threshold: float,
        iou_threshold: float,
        mode: str,
        predictor: Predictor,
    ) -> np.ndarray:
        # input is rgb image, output must be rgb too
        predictor.detector.detector_kwargs["conf"] = score_threshold
        predictor.detector.detector_kwargs["iou"] = iou_threshold
        if mode == "Use persons and faces":
            use_persons = True
            disable_faces = False

        elif mode == "Use persons only":
            use_persons = True
            disable_faces = True

        elif mode == "Use faces only":
            use_persons = False
            disable_faces = False

        predictor.age_gender_model.meta.use_persons = use_persons
        predictor.age_gender_model.meta.disable_faces = disable_faces
        # image = image[:, :, ::-1]  # RGB -> BGR
        detected_objects, out_im = predictor.recognize(image)
        has_child, has_female, has_male = False, False, False
        if len(detected_objects.ages) > 0:
            has_child = min(detected_objects.ages) < 18
            has_female = "female" in detected_objects.genders
            has_male = "male" in detected_objects.genders

        return out_im[:, :, ::-1], has_child, has_female, has_male

    def valid_img(self, img_path):
        image = cv2.imread(img_path)
        return self._detect(image, 0.4, 0.7, "Use persons and faces", self.predictor)


def infer(photo: str):
    if is_url(photo):
        if os.path.exists(TMP_DIR):
            shutil.rmtree(TMP_DIR)

        photo = download_file(photo, f"{TMP_DIR}/download.jpg")

    detector = ValidImgDetector()
    if not photo or not os.path.exists(photo) or imghdr.what(photo) == None:
        return None, None, None, "请正确输入图片 Please input image correctly"

    return detector.valid_img(photo)


if __name__ == "__main__":
    with gr.Blocks() as iface:
        warnings.filterwarnings("ignore")
        with gr.Tab("上传模式 Upload Mode"):
            gr.Interface(
                fn=infer,
                inputs=gr.Image(label="上传照片 Upload Photo", type="filepath"),
                outputs=[
                    gr.Image(label="检测结果 Detection Result", type="numpy"),
                    gr.Textbox(label="存在儿童 Has Child"),
                    gr.Textbox(label="存在女性 Has Female"),
                    gr.Textbox(label="存在男性 Has Male"),
                ],
                examples=get_jpg_files(f"{MODEL_DIR}/examples"),
                allow_flagging="never",
            )

        with gr.Tab("在线模式 Online Mode"):
            gr.Interface(
                fn=infer,
                inputs=gr.Textbox(label="网络图片链接 Online Picture URL"),
                outputs=[
                    gr.Image(label="检测结果 Detection Result", type="numpy"),
                    gr.Textbox(label="存在儿童 Has Child"),
                    gr.Textbox(label="存在女性 Has Female"),
                    gr.Textbox(label="存在男性 Has Male"),
                ],
                allow_flagging="never",
                cache_examples=False,
            )

    iface.launch()