File size: 5,294 Bytes
2de1f98
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import os
import sys
import os.path as osp
from pathlib import Path
import cv2
import gradio as gr
import torch
import math

try:
    import mmpose
except:
    os.system('pip install /home/user/app/main/transformer_utils')

os.system('cp -rf /home/user/app/assets/conversions.py /home/user/.pyenv/versions/3.8.18/lib/python3.8/site-packages/torchgeometry/core/conversions.py')
DEFAULT_MODEL='smpler_x_h32'
OUT_FOLDER = '/home/user/app/demo_out'
os.makedirs(OUT_FOLDER, exist_ok=True)
num_gpus = 1 if torch.cuda.is_available() else -1
print("!!!", torch.cuda.is_available())      
print(torch.cuda.device_count())   
print(torch.version.cuda)  
index = torch.cuda.current_device()
print(index)  
print(torch.cuda.get_device_name(index))
from main.inference import Inferer
inferer = Inferer(DEFAULT_MODEL, num_gpus, OUT_FOLDER)

def infer(video_input, in_threshold=0.5, num_people="Single person", render_mesh=False):
    os.system(f'rm -rf {OUT_FOLDER}/*')
    multi_person = False if (num_people == "Single person") else True
    cap = cv2.VideoCapture(video_input)
    fps = math.ceil(cap.get(5))
    width = int(cap.get(3))
    height = int(cap.get(4))
    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
    video_path = osp.join(OUT_FOLDER, f'out.m4v')
    final_video_path = osp.join(OUT_FOLDER, f'out.mp4')
    video_output = cv2.VideoWriter(video_path, fourcc, fps, (width, height))
    success = 1
    frame = 0
    while success:
        success, original_img = cap.read()
        if not success:
            break
        frame += 1
        img, mesh_paths, smplx_paths = inferer.infer(original_img, in_threshold, frame, multi_person, not(render_mesh))
        video_output.write(img)
    cap.release()
    video_output.release()
    cv2.destroyAllWindows()
    os.system(f'ffmpeg -i {video_path} -c copy {final_video_path}')

    #Compress mesh and smplx files
    save_path_mesh = os.path.join(OUT_FOLDER, 'mesh')
    save_mesh_file = os.path.join(OUT_FOLDER, 'mesh.zip')
    os.makedirs(save_path_mesh, exist_ok= True)
    save_path_smplx = os.path.join(OUT_FOLDER, 'smplx')
    save_smplx_file = os.path.join(OUT_FOLDER, 'smplx.zip')
    os.makedirs(save_path_smplx, exist_ok= True)
    os.system(f'zip -r {save_mesh_file} {save_path_mesh}')
    os.system(f'zip -r {save_smplx_file} {save_path_smplx}')
    return video_path, save_mesh_file, save_smplx_file

TITLE = '''<h1 align="center">SMPLer-X: Scaling Up Expressive Human Pose and Shape Estimation</h1>'''
VIDEO = '''
<center><iframe width="960" height="540" 
src="https://www.youtube.com/embed/DepTqbPpVzY?si=qSeQuX-bgm_rON7E"title="SMPLer-X: Scaling Up Expressive Human Pose and Shape Estimation" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen>
</iframe>
</center><br>'''
DESCRIPTION = '''
<b>Official Gradio demo</b> for <a href="https://caizhongang.com/projects/SMPLer-X/"><b>SMPLer-X: Scaling Up Expressive Human Pose and Shape Estimation</b></a>.<br>
<p>
Note: You can drop a video at the panel (or select one of the examples) 
    then you will get the 3D reconstructions of the detected human. ).
</p>
'''

with gr.Blocks(title="SMPLer-X", css=".gradio-container") as demo:

    gr.Markdown(TITLE)
    gr.HTML(VIDEO)
    gr.Markdown(DESCRIPTION)

    with gr.Row():
        with gr.Column():
            video_input = gr.Video(label="Input video", elem_classes="video")
            threshold = gr.Slider(0, 1.0, value=0.5, label='BBox detection threshold')
        with gr.Column(scale=2):
            num_people = gr.Radio(
                choices=["Single person", "Multiple people"],
                value="Single person",
                label="Number of people",
                info="Choose how many people are there in the video. Choose 'single person' for faster inference.",
                interactive=True,
                scale=1,)
            gr.HTML("""<br/>""")
            mesh_as_vertices = gr.Checkbox(
                label="Render mesh", 
                info="By default, the reconstructions of human bodies are expressed as vertices for faster inference. Check this option if you want to render the human body with mesh.",
                interactive=True,
                scale=1,)

            send_button = gr.Button("Infer")
    gr.HTML("""<br/>""")

    with gr.Row():
        with gr.Column():
            video_output = gr.Video(elem_classes="video")
        with gr.Column():
            meshes_output = gr.File(label="3D meshes")
            smplx_output = gr.File(label= "SMPL-X models")
    # example_images = gr.Examples([])
    send_button.click(fn=infer, inputs=[video_input, threshold, num_people, mesh_as_vertices], outputs=[video_output, meshes_output, smplx_output])
    # with gr.Row():
    example_videos = gr.Examples([
        ['/home/user/app/assets/01.mp4'], 
        ['/home/user/app/assets/02.mp4'], 
        ['/home/user/app/assets/03.mp4'],
        ['/home/user/app/assets/04.mp4'],
        ['/home/user/app/assets/05.mp4'], 
        ['/home/user/app/assets/06.mp4'], 
        ['/home/user/app/assets/07.mp4'],
        ['/home/user/app/assets/08.mp4'],
        ['/home/user/app/assets/09.mp4'],
        ], 
        inputs=[video_input, 0.5])

#demo.queue()
demo.launch(debug=True)