File size: 4,104 Bytes
54b1993
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7074d41
54b1993
072bc06
54b1993
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
072bc06
 
 
54b1993
 
df24eff
 
 
 
 
 
 
 
 
 
 
 
ac37187
df24eff
 
 
 
 
 
9dbf189
df24eff
 
 
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
#!/usr/bin/env python

import functools
import os
import pathlib

import cv2
import dlib
import gradio as gr
import huggingface_hub
import numpy as np
import pretrainedmodels
import torch
import torch.nn as nn
import torch.nn.functional as F

DESCRIPTION = '# [Age Estimation](https://github.com/yu4u/age-estimation-pytorch)'


def get_model(model_name='se_resnext50_32x4d',
              num_classes=101,
              pretrained='imagenet'):
    model = pretrainedmodels.__dict__[model_name](pretrained=pretrained)
    dim_feats = model.last_linear.in_features
    model.last_linear = nn.Linear(dim_feats, num_classes)
    model.avg_pool = nn.AdaptiveAvgPool2d(1)
    return model


def load_model(device):
    model = get_model(model_name='se_resnext50_32x4d', pretrained=None)
    path = huggingface_hub.hf_hub_download(
        'public-data/yu4u-age-estimation-pytorch', 'pretrained.pth')
    model.load_state_dict(torch.load(path))
    model = model.to(device)
    model.eval()
    return model


def load_image(path):
    image = cv2.imread(path)
    h_orig, w_orig = image.shape[:2]
    size = max(h_orig, w_orig)
    scale = 640 / size
    w, h = int(w_orig * scale), int(h_orig * scale)
    image = cv2.resize(image, (w, h))
    return image


def draw_label(image,
               point,
               label,
               font=cv2.FONT_HERSHEY_SIMPLEX,
               font_scale=0.8,
               thickness=1):
    size = cv2.getTextSize(label, font, font_scale, thickness)[0]
    x, y = point
    cv2.rectangle(image, (x, y - size[1]), (x + size[0], y), (255, 0, 0),
                  cv2.FILLED)
    cv2.putText(image,
                label,
                point,
                font,
                font_scale, (255, 255, 255),
                thickness,
                lineType=cv2.LINE_AA)


@torch.inference_mode()
def predict(image, model, face_detector, device, margin=0.4, input_size=224):
    image = cv2.imread(image, cv2.IMREAD_COLOR)[:, :, ::-1].copy()
    image_h, image_w = image.shape[:2]

    # detect faces using dlib detector
    detected = face_detector(image, 3)
    faces = np.empty((len(detected), input_size, input_size, 3))
    age_data = []

    if len(detected) > 0:
        for i, d in enumerate(detected):
            x1, y1, x2, y2, w, h = d.left(), d.top(
            ), d.right() + 1, d.bottom() + 1, d.width(), d.height()
            xw1 = max(int(x1 - margin * w), 0)
            yw1 = max(int(y1 - margin * h), 0)
            xw2 = min(int(x2 + margin * w), image_w - 1)
            yw2 = min(int(y2 + margin * h), image_h - 1)
            faces[i] = cv2.resize(image[yw1:yw2 + 1, xw1:xw2 + 1],
                                  (input_size, input_size))

            cv2.rectangle(image, (x1, y1), (x2, y2), (255, 255, 255), 2)
            cv2.rectangle(image, (xw1, yw1), (xw2, yw2), (255, 0, 0), 2)

        # predict ages
        inputs = torch.from_numpy(
            np.transpose(faces.astype(np.float32), (0, 3, 1, 2))).to(device)
        outputs = F.softmax(model(inputs), dim=-1).cpu().numpy()
        ages = np.arange(0, 101)
        predicted_ages = (outputs * ages).sum(axis=-1)

        # draw results
        for age, d in zip(predicted_ages, detected):
            age_text = f'{int(age)}'
            age_data.append({'age': int(age), 'text': age_text, 'face_coordinates': (d.left(), d.top())})
    return age_data


def main():
    device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
    model = load_model(device)
    face_detector = dlib.get_frontal_face_detector()
    fn = functools.partial(predict, model=model, face_detector=face_detector, device=device)

    image_dir = pathlib.Path('sample_images')
    examples = [path.as_posix() for path in sorted(image_dir.glob('*.jpg'))]

    demo = gr.Interface(
        fn=fn,
        inputs=gr.inputs.Image(type="filepath"),
        outputs="json",
        examples=examples,
        title="Age Estimation",
        description=DESCRIPTION,
        cache_examples=os.getenv('CACHE_EXAMPLES') == '1'
    )

    demo.launch()

if __name__ == '__main__':
    main()