File size: 5,680 Bytes
42b0b31
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
# Self-Correction-Human-Parsing
# Original https://github.com/GoGoDuck912/Self-Correction-Human-Parsing

import os
import torch
import numpy as np
from PIL import Image
import cv2

import torchvision.transforms as T

from .transforms import transform_logits, get_affine_transform
from . import networks
from annotator.util import annotator_ckpts_path
from huggingface_hub import snapshot_download

dataset_settings = {
    'lip': {
        'input_size': [473, 473],
        'num_classes': 20,
        'label': ['Background', 'Hat', 'Hair', 'Glove', 'Sunglasses', 'Upper-clothes', 'Dress', 'Coat',
                  'Socks', 'Pants', 'Jumpsuits', 'Scarf', 'Skirt', 'Face', 'Left-arm', 'Right-arm',
                  'Left-leg', 'Right-leg', 'Left-shoe', 'Right-shoe']
    },
    'atr': {
        'input_size': [512, 512],
        'num_classes': 18,
        'label': ['Background', 'Hat', 'Hair', 'Sunglasses', 'Upper-clothes', 'Skirt', 'Pants', 'Dress', 'Belt',
                  'Left-shoe', 'Right-shoe', 'Face', 'Left-leg', 'Right-leg', 'Left-arm', 'Right-arm', 'Bag', 'Scarf']
    },
    'pascal': {
        'input_size': [512, 512],
        'num_classes': 7,
        'label': ['Background', 'Head', 'Torso', 'Upper Arms', 'Lower Arms', 'Upper Legs', 'Lower Legs'],
    }
}


def get_palette(num_cls):
    """ Returns the color map for visualizing the segmentation mask.
    Args:
        num_cls: Number of classes
    Returns:
        The color map
    """
    n = num_cls
    palette = [0] * (n * 3)
    for j in range(0, n):
        lab = j
        palette[j * 3 + 0] = 0
        palette[j * 3 + 1] = 0
        palette[j * 3 + 2] = 0
        i = 0
        while lab:
            palette[j * 3 + 0] |= (((lab >> 0) & 1) << (7 - i))
            palette[j * 3 + 1] |= (((lab >> 1) & 1) << (7 - i))
            palette[j * 3 + 2] |= (((lab >> 2) & 1) << (7 - i))
            i += 1
            lab >>= 3
    return palette

class Segmentator(torch.nn.Module):
    def __init__(self, dataset='lip'):
        super().__init__()

        num_classes = dataset_settings[dataset]['num_classes']
        input_size = dataset_settings[dataset]['input_size']
        label = dataset_settings[dataset]['label']

        if dataset == 'atr':
            model_path='exp-schp-201908301523-atr.pth'
        elif dataset == 'lip':
            model_path='exp-schp-201908261155-lip.pth'
        
        model_path = os.path.join(annotator_ckpts_path, model_path)

        snapshot_download(repo_id="soonyau/visconet", allow_patterns="exp-schp-201908301523-atr.pth", local_dir=annotator_ckpts_path)

        self.model = networks.init_model('resnet101', num_classes=num_classes, pretrained=None)
        state_dict = torch.load(model_path)['state_dict']
        from collections import OrderedDict
        new_state_dict = OrderedDict()
        for k, v in state_dict.items():
            name = k[7:]  # remove `module.`
            new_state_dict[name] = v
        self.model.load_state_dict(new_state_dict)
        self.model.eval()

        self.palette = get_palette(num_classes)

        self.transform = T.Compose([
            T.ToTensor(),
            T.Normalize(mean=[0.406, 0.456, 0.485], std=[0.225, 0.224, 0.229])
        ])
        self.aspect_ratio = input_size[1] * 1.0 / input_size[0]
        self.input_size = np.asarray(input_size)

    def _box2cs(self, box):
        x, y, w, h = box[:4]
        return self._xywh2cs(x, y, w, h)

    def _xywh2cs(self, x, y, w, h):
        center = np.zeros((2), dtype=np.float32)
        center[0] = x + w * 0.5
        center[1] = y + h * 0.5
        if w > self.aspect_ratio * h:
            h = w * 1.0 / self.aspect_ratio
        elif w < self.aspect_ratio * h:
            w = h * self.aspect_ratio
        scale = np.array([w, h], dtype=np.float32)
        return center, scale
    
    def preprocess(self, image:np.array):
        # convert numpy to cv2
        image = image[:,:,::-1]
        h, w, _ = image.shape

        # Get person center and scale
        person_center, s = self._box2cs([0, 0, w - 1, h - 1])
        r = 0
        trans = get_affine_transform(person_center, s, r, self.input_size)
        input = cv2.warpAffine(
            image,
            trans,
            (int(self.input_size[1]), int(self.input_size[0])),
            flags=cv2.INTER_LINEAR,
            borderMode=cv2.BORDER_CONSTANT,
            borderValue=(0, 0, 0))

        input = self.transform(input)
        meta = {
            'center': person_center,
            'height': h,
            'width': w,
            'scale': s,
            'rotation': r
        }

        return input, meta

    @torch.no_grad()
    def __call__(self, input_image):
        image, meta = self.preprocess(input_image)
        c = meta['center']
        s = meta['scale']
        w = meta['width']
        h = meta['height']
        input_size = list(self.input_size)
        device = next(self.parameters()).device
        output = self.model(image.unsqueeze(0).to(device))
        upsample = torch.nn.Upsample(size=input_size, mode='bilinear', align_corners=True)
        upsample_output = upsample(output[0][-1][0].unsqueeze(0))
        upsample_output = upsample_output.squeeze()
        upsample_output = upsample_output.permute(1, 2, 0)  # CHW -> HWC
        logits_result = transform_logits(upsample_output.data.cpu().numpy(), c, s, w, h, input_size=input_size)
        parsing_result = np.argmax(logits_result, axis=2)
        output_img = Image.fromarray(np.asarray(parsing_result, dtype=np.uint8))
        #return output_img
        output_img.putpalette(self.palette)
        return output_img
        #return np.array(output_img)