File size: 4,746 Bytes
5c22f66
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
#!/usr/bin/env python

from __future__ import annotations

import argparse
import os
import pathlib
import subprocess
import tarfile

if os.environ.get('SYSTEM') == 'spaces':
    subprocess.call('pip uninstall -y opencv-python'.split())
    subprocess.call('pip uninstall -y opencv-python-headless'.split())
    subprocess.call('pip install opencv-python-headless==4.5.5.64'.split())

import gradio as gr
import huggingface_hub
import mediapipe as mp
import numpy as np

mp_drawing = mp.solutions.drawing_utils
mp_drawing_styles = mp.solutions.drawing_styles
mp_pose = mp.solutions.pose

TITLE = 'MediaPipe Human Pose Estimation'
DESCRIPTION = 'https://google.github.io/mediapipe/'
ARTICLE = None

TOKEN = os.environ['TOKEN']


def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser()
    parser.add_argument('--theme', type=str)
    parser.add_argument('--live', action='store_true')
    parser.add_argument('--share', action='store_true')
    parser.add_argument('--port', type=int)
    parser.add_argument('--disable-queue',
                        dest='enable_queue',
                        action='store_false')
    parser.add_argument('--allow-flagging', type=str, default='never')
    parser.add_argument('--allow-screenshot', action='store_true')
    return parser.parse_args()


def load_sample_images() -> list[pathlib.Path]:
    image_dir = pathlib.Path('images')
    if not image_dir.exists():
        image_dir.mkdir()
        dataset_repo = 'hysts/input-images'
        filenames = ['002.tar']
        for name in filenames:
            path = huggingface_hub.hf_hub_download(dataset_repo,
                                                   name,
                                                   repo_type='dataset',
                                                   use_auth_token=TOKEN)
            with tarfile.open(path) as f:
                f.extractall(image_dir.as_posix())
    return sorted(image_dir.rglob('*.jpg'))


def run(image: np.ndarray, model_complexity: int, enable_segmentation: bool,
        min_detection_confidence: float, background_color: str) -> np.ndarray:
    with mp_pose.Pose(
            static_image_mode=True,
            model_complexity=model_complexity,
            enable_segmentation=enable_segmentation,
            min_detection_confidence=min_detection_confidence) as pose:
        results = pose.process(image)

    res = image[:, :, ::-1].copy()
    if enable_segmentation:
        if background_color == 'white':
            bg_color = 255
        elif background_color == 'black':
            bg_color = 0
        elif background_color == 'green':
            bg_color = (0, 255, 0)
        else:
            raise ValueError

        if results.segmentation_mask is not None:
            res[results.segmentation_mask <= 0.1] = bg_color
        else:
            res[:] = bg_color

    mp_drawing.draw_landmarks(res,
                              results.pose_landmarks,
                              mp_pose.POSE_CONNECTIONS,
                              landmark_drawing_spec=mp_drawing_styles.
                              get_default_pose_landmarks_style())

    return res[:, :, ::-1]


def main():
    args = parse_args()

    model_complexities = list(range(3))
    background_colors = ['white', 'black', 'green']

    image_paths = load_sample_images()
    examples = [[
        path.as_posix(), model_complexities[1], True, 0.5, background_colors[0]
    ] for path in image_paths]

    gr.Interface(
        run,
        [
            gr.inputs.Image(type='numpy', label='Input'),
            gr.inputs.Radio(model_complexities,
                            type='index',
                            default=model_complexities[1],
                            label='Model Complexity'),
            gr.inputs.Checkbox(default=True, label='Enable Segmentation'),
            gr.inputs.Slider(0,
                             1,
                             step=0.05,
                             default=0.5,
                             label='Minimum Detection Confidence'),
            gr.inputs.Radio(background_colors,
                            type='value',
                            default=background_colors[0],
                            label='Background Color'),
        ],
        gr.outputs.Image(type='numpy', label='Output'),
        examples=examples,
        title=TITLE,
        description=DESCRIPTION,
        article=ARTICLE,
        theme=args.theme,
        allow_screenshot=args.allow_screenshot,
        allow_flagging=args.allow_flagging,
        live=args.live,
    ).launch(
        enable_queue=args.enable_queue,
        server_port=args.port,
        share=args.share,
    )


if __name__ == '__main__':
    main()