app.py

# Face Recognition Hub
# author: Zeng Yifu（曾逸夫）
# creation time: 2022-07-28
# email: [email protected]
# project homepage: https://gitee.com/CV_Lab/face-recognition-hub

import os
import sys
from pathlib import Path

import face_recognition
import gradio as gr
from PIL import Image, ImageDraw, ImageFont

from util.fonts_opt import is_fonts

ROOT_PATH = sys.path[0]  # 项目根目录

IMG_PATH_Test = "./img_examples/unknown"

FONTSIZE = 15

OCR_TR_DESCRIPTION = '''# Face Recognition
<div id="content_align">https://github.com/ageitgey/face_recognition demo</div>'''

def str_intercept(img_path):
    img_path_ = img_path[::-1]
    point_index = 0  # 记录反转后第一个点的位置
    slash_index = 0  # 记录反转后第一个斜杠的位置

    flag_pi = 0
    flag_si = 0

    for i in range(len(img_path_)):
        if (img_path_[i] == "." and flag_pi == 0):
            point_index = i
            flag_pi = 1

        if (img_path_[i] == "/" and flag_si == 0):
            slash_index = i
            flag_si = 1

    point_index = len(img_path) - 1 - point_index
    slash_index = len(img_path) - 1 - slash_index

    return point_index, slash_index


# 人脸录入
def face_entry(img_path, name_text):
    if img_path == "" or name_text == "" or img_path is None or name_text is None:
        return None, None, None

    point_index, slash_index = str_intercept(img_path)
    img_renamePath = f"{img_path[:slash_index+1]}{name_text}{img_path[point_index:]}"
    os.rename(img_path, img_renamePath)
    img_ = Image.open(img_renamePath)
    print(img_renamePath)

    return img_, img_renamePath, name_text


# 设置示例
def set_example_image(example: list):
    return gr.Image.update(value=example[0])


def face_recognition_(img_srcPath, img_tagPath, img_personName):
    if img_tagPath == "" or img_tagPath is None:
        return None

    image_of_person = face_recognition.load_image_file(img_srcPath)
    person_face_encoding = face_recognition.face_encodings(image_of_person)[0]

    known_face_encodings = [
        person_face_encoding,]

    known_face_names = [
        img_personName,]

    test_image = face_recognition.load_image_file(img_tagPath)

    face_locations = face_recognition.face_locations(test_image)
    face_encodings = face_recognition.face_encodings(test_image, face_locations)

    pil_image = Image.fromarray(test_image)
    img_pil = ImageDraw.Draw(pil_image)
    textFont = ImageFont.truetype(str(f"{ROOT_PATH}/fonts/SimSun.ttf"), size=FONTSIZE)
    #     ymin, xmax, ymax, xmin
    for (top, right, bottom, left), face_encoding in zip(face_locations, face_encodings):
        matches = face_recognition.compare_faces(known_face_encodings, face_encoding)

        name = "Unknown Person"

        if True in matches:
            first_matches_index = matches.index(True)
            name = known_face_names[first_matches_index]

        img_pil.rectangle([left, top, right, bottom], fill=None, outline=(255, 228, 181), width=2)  # 边界框
        text_w, text_h = textFont.getsize(name)  # 标签尺寸
        # 标签背景
        img_pil.rectangle(
            (left, top, left + text_w, top + text_h),
            fill=(255, 255, 255),
            outline=(255, 255, 255),
        )

        # 标签
        img_pil.multiline_text(
            (left, top),
            name,
            fill=(0, 0, 0),
            font=textFont,
            align="center",
        )

    del img_pil
    return pil_image


def main():
    is_fonts(f"{ROOT_PATH}/fonts")  # 检查字体文件

    with gr.Blocks(css='style.css') as demo:
        gr.Markdown(OCR_TR_DESCRIPTION)

        # -------------- 人脸识别 录入 --------------
        with gr.Row():
            gr.Markdown("### Step 01: Face Entry")
        with gr.Row():
            with gr.Column():
                with gr.Row():
                    input_img = gr.Image(image_mode="RGB", source="upload", type="filepath", label="face entry")
                with gr.Row():
                    input_name = gr.Textbox(label="Name")
                with gr.Row():
                    btn = gr.Button(value="Entry")

            with gr.Column():
                with gr.Row():
                    output_ = gr.Image(image_mode="RGB", source="upload", type="pil", label="entry image")
                    input_srcImg = gr.Variable(value="")
                    input_srcName = gr.Variable(value="")
        with gr.Row():
            example_list = [["./img_examples/known/ChengLong.jpg", "成龙"],
                            ["./img_examples/known/VinDiesel.jpg", "VinDiesel"],
                            ["./img_examples/known/JasonStatham.jpg", "JasonStatham"],
                            ["./img_examples/known/ZhenZidan.jpg", "甄子丹"]]
            gr.Examples(example_list,
                        [input_img, input_name],
                        output_,
                        set_example_image,
                        cache_examples=False)


        # -------------- 人脸识别 测试 --------------
        with gr.Row():
                gr.Markdown("### Step 02: Face Test")
        with gr.Row():
            with gr.Column():
                with gr.Row():
                    input_img_test = gr.Image(image_mode="RGB", source="upload", type="filepath", label="test image")
                with gr.Row():
                    btn_test = gr.Button(value="Test")
                with gr.Row():
                    paths = sorted(Path(IMG_PATH_Test).rglob('*.jpg'))
                    example_images_test = gr.Dataset(components=[input_img],
                                                     samples=[[path.as_posix()] for path in paths])

            with gr.Column():
                with gr.Row():
                    output_test = gr.Image(image_mode="RGB", source="upload", type="pil", label="identify image")

        btn.click(fn=face_entry, inputs=[input_img, input_name], outputs=[output_, input_srcImg, input_srcName])

        btn_test.click(fn=face_recognition_,
                       inputs=[input_srcImg, input_img_test, input_srcName],
                       outputs=[output_test])
        example_images_test.click(fn=set_example_image, inputs=[
            example_images_test,], outputs=[
                input_img_test,])

        return demo


if __name__ == "__main__":
    demo = main()
    demo.launch(inbrowser=True)