-
Notifications
You must be signed in to change notification settings - Fork 13
/
Copy pathsegformer_b3_fashion.py
265 lines (246 loc) · 11.9 KB
/
segformer_b3_fashion.py
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
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
import os
import numpy as np
from urllib.request import urlopen
import torchvision.transforms as transforms
import folder_paths
from transformers import SegformerImageProcessor, AutoModelForSemanticSegmentation
from PIL import Image,ImageOps, ImageFilter
import torch.nn as nn
import torch
# comfy_path = os.path.dirname(folder_paths.__file__)
# custom_nodes_path = os.path.join(comfy_path, "custom_nodes")
# 指定本地分割模型文件夹的路径
# model_folder_path = os.path.join(custom_nodes_path,"Comfyui_segformer_b2_clothes","checkpoints","segformer_b3_fashion")
model_folder_path = os.path.join(folder_paths.models_dir,"segformer_b3_fashion")
processor = SegformerImageProcessor.from_pretrained(model_folder_path)
model = AutoModelForSemanticSegmentation.from_pretrained(model_folder_path)
def tensor2pil(image):
return Image.fromarray(np.clip(255. * image.cpu().numpy().squeeze(), 0, 255).astype(np.uint8))
# Convert PIL to Tensor
def pil2tensor(image):
return torch.from_numpy(np.array(image).astype(np.float32) / 255.0).unsqueeze(0)
# 切割服装
def get_segmentation(tensor_image):
cloth = tensor2pil(tensor_image)
# 预处理和预测
inputs = processor(images=cloth, return_tensors="pt")
outputs = model(**inputs)
logits = outputs.logits.cpu()
upsampled_logits = nn.functional.interpolate(logits, size=cloth.size[::-1], mode="bilinear", align_corners=False)
pred_seg = upsampled_logits.argmax(dim=1)[0].numpy()
return pred_seg,cloth
class segformer_b3_fashion:
def __init__(self):
pass
# Labels: 0: "Background", 1: "Hat", 2: "Hair", 3: "Sunglasses", 4: "Upper-clothes", 5: "Skirt", 6: "Pants", 7: "Dress", 8: "Belt", 9: "Left-shoe", 10: "Right-shoe", 11: "Face", 12: "Left-leg", 13: "Right-leg", 14: "Left-arm", 15: "Right-arm", 16: "Bag", 17: "Scarf"
@classmethod
def INPUT_TYPES(cls):
return {"required":
{
"image":("IMAGE", {"default": "","multiline": False}),
"shirt": ("BOOLEAN", {"default": False, "label_on": "enabled(衬衫、罩衫)", "label_off": "disabled(衬衫、罩衫)"}),
"top": ("BOOLEAN", {"default": False, "label_on": "enabled(上衣、t恤)", "label_off": "disabled(上衣、t恤)"}),
"sweater": ("BOOLEAN", {"default": False, "label_on": "enabled(毛衣)", "label_off": "disabled(毛衣)"}),
"cardigan": ("BOOLEAN", {"default": False, "label_on": "enabled(开襟羊毛衫)", "label_off": "disabled(开襟羊毛衫)"}),
"jacket": ("BOOLEAN", {"default": False, "label_on": "enabled(夹克)", "label_off": "disabled(夹克)"}),
"vest": ("BOOLEAN", {"default": False, "label_on": "enabled(背心)", "label_off": "disabled(背心)"}),
"pants": ("BOOLEAN", {"default": False, "label_on": "enabled(裤子)", "label_off": "disabled(裤子)"}),
"shorts": ("BOOLEAN", {"default": False, "label_on": "enabled(短裤)", "label_off": "disabled(短裤)"}),
"skirt": ("BOOLEAN", {"default": False, "label_on": "enabled(裙子)", "label_off": "disabled(裙子)"}),
"coat": ("BOOLEAN", {"default": False, "label_on": "enabled(外套)", "label_off": "disabled(外套)"}),
"dress": ("BOOLEAN", {"default": False, "label_on": "enabled(连衣裙)", "label_off": "disabled(连衣裙)"}),
"jumpsuit": ("BOOLEAN", {"default": False, "label_on": "enabled(连身裤)", "label_off": "disabled(连身裤)"}),
"cape": ("BOOLEAN", {"default": False, "label_on": "enabled(斗篷)", "label_off": "disabled(斗篷)"}),
"glasses": ("BOOLEAN", {"default": False, "label_on": "enabled(眼镜)", "label_off": "disabled(眼镜)"}),
"hat": ("BOOLEAN", {"default": False, "label_on": "enabled(帽子)", "label_off": "disabled(帽子)"}),
"hairaccessory": ("BOOLEAN", {"default": False, "label_on": "enabled(头带)", "label_off": "disabled(头带)"}),
"tie": ("BOOLEAN", {"default": False, "label_on": "enabled(领带)", "label_off": "disabled(领带)"}),
"glove": ("BOOLEAN", {"default": False, "label_on": "enabled(手套)", "label_off": "disabled(手套)"}),
"watch": ("BOOLEAN", {"default": False, "label_on": "enabled(手表)", "label_off": "disabled(手表)"}),
"belt": ("BOOLEAN", {"default": False, "label_on": "enabled(皮带)", "label_off": "disabled(皮带)"}),
"legwarmer": ("BOOLEAN", {"default": False, "label_on": "enabled(暖腿器)", "label_off": "disabled(暖腿器)"}),
"tights": ("BOOLEAN", {"default": False, "label_on": "enabled(紧身衣、长筒袜)", "label_off": "disabled(紧身衣、长筒袜)"}),
"sock": ("BOOLEAN", {"default": False, "label_on": "enabled(袜子)", "label_off": "disabled(袜子)"}),
"shoe": ("BOOLEAN", {"default": False, "label_on": "enabled(鞋子)", "label_off": "disabled(鞋子)"}),
"bagwallet": ("BOOLEAN", {"default": False, "label_on": "enabled(包、钱包)", "label_off": "disabled(包、钱包)"}),
"scarf": ("BOOLEAN", {"default": False, "label_on": "enabled(围巾)", "label_off": "disabled(围巾)"}),
"umbrella": ("BOOLEAN", {"default": False, "label_on": "enabled(雨伞)", "label_off": "disabled(雨伞)"}),
"hood": ("BOOLEAN", {"default": False, "label_on": "enabled(兜帽)", "label_off": "disabled(兜帽)"}),
"collar": ("BOOLEAN", {"default": False, "label_on": "enabled(衣领)", "label_off": "disabled(衣领)"}),
"lapel": ("BOOLEAN", {"default": False, "label_on": "enabled(翻领)", "label_off": "disabled(翻领)"}),
"epaulette": ("BOOLEAN", {"default": False, "label_on": "enabled(肩章)", "label_off": "disabled(肩章)"}),
"sleeve": ("BOOLEAN", {"default": False, "label_on": "enabled(袖子)", "label_off": "disabled(袖子)"}),
"pocket": ("BOOLEAN", {"default": False, "label_on": "enabled(口袋)", "label_off": "disabled(口袋)"}),
"neckline": ("BOOLEAN", {"default": False, "label_on": "enabled(领口)", "label_off": "disabled(领口)"}),
"buckle": ("BOOLEAN", {"default": False, "label_on": "enabled(带扣)", "label_off": "disabled(带扣)"}),
"zipper": ("BOOLEAN", {"default": False, "label_on": "enabled(拉链)", "label_off": "disabled(拉链)"}),
"applique": ("BOOLEAN", {"default": False, "label_on": "enabled(贴花)", "label_off": "disabled(贴花)"}),
"bead": ("BOOLEAN", {"default": False, "label_on": "enabled(珠子)", "label_off": "disabled(珠子)"}),
"bow": ("BOOLEAN", {"default": False, "label_on": "enabled(蝴蝶结)", "label_off": "disabled(蝴蝶结)"}),
"flower": ("BOOLEAN", {"default": False, "label_on": "enabled(花)", "label_off": "disabled(花)"}),
"fringe": ("BOOLEAN", {"default": False, "label_on": "enabled(额前短垂发)", "label_off": "disabled(额前短垂发)"}),
"ribbon": ("BOOLEAN", {"default": False, "label_on": "enabled(丝带)", "label_off": "disabled(丝带)"}),
"rivet": ("BOOLEAN", {"default": False, "label_on": "enabled(铆钉)", "label_off": "disabled(铆钉)"}),
"ruffle": ("BOOLEAN", {"default": False, "label_on": "enabled(褶饰)", "label_off": "disabled(褶饰)"}),
"sequin": ("BOOLEAN", {"default": False, "label_on": "enabled(亮片)", "label_off": "disabled(亮片)"}),
"tassel": ("BOOLEAN", {"default": False, "label_on": "enabled(流苏)", "label_off": "disabled(流苏)"}),
}
}
RETURN_TYPES = ("IMAGE",)
RETURN_NAMES = ("mask_image",)
OUTPUT_NODE = True
FUNCTION = "sample"
CATEGORY = "CXH"
def sample(self,image,
shirt,
top,
sweater,
cardigan,
jacket,
vest,
pants,
shorts,
skirt,
coat,
dress,
jumpsuit,
cape,
glasses,
hat,
hairaccessory,
tie,
glove,
watch,
belt,
legwarmer,
tights,
sock,
shoe,
bagwallet,
scarf,
umbrella,
hood,
collar,
lapel,
epaulette,
sleeve,
pocket,
neckline,
buckle,
zipper,
applique,
bead,
bow,
flower,
fringe,
ribbon,
rivet,
ruffle,
sequin,
tassel):
results = []
for item in image:
# seg切割结果，衣服pil
pred_seg,cloth = get_segmentation(item)
labels_to_keep = [0]
# if background :
# labels_to_keep.append(0)
if not shirt:
labels_to_keep.append(1)
if not top:
labels_to_keep.append(2)
if not sweater:
labels_to_keep.append(3)
if not cardigan:
labels_to_keep.append(4)
if not jacket:
labels_to_keep.append(5)
if not vest:
labels_to_keep.append(6)
if not pants:
labels_to_keep.append(7)
if not shorts:
labels_to_keep.append(8)
if not skirt:
labels_to_keep.append(9)
if not coat:
labels_to_keep.append(10)
if not dress:
labels_to_keep.append(11)
if not jumpsuit:
labels_to_keep.append(12)
if not cape:
labels_to_keep.append(13)
if not glasses:
labels_to_keep.append(14)
if not hat:
labels_to_keep.append(15)
if not hairaccessory:
labels_to_keep.append(16)
if not tie:
labels_to_keep.append(17)
if not glove:
labels_to_keep.append(18)
if not watch:
labels_to_keep.append(19)
if not belt:
labels_to_keep.append(20)
if not legwarmer:
labels_to_keep.append(21)
if not tights:
labels_to_keep.append(22)
if not sock:
labels_to_keep.append(23)
if not shoe:
labels_to_keep.append(24)
if not bagwallet:
labels_to_keep.append(25)
if not scarf:
labels_to_keep.append(26)
if not umbrella:
labels_to_keep.append(27)
if not hood:
labels_to_keep.append(28)
if not collar:
labels_to_keep.append(29)
if not lapel:
labels_to_keep.append(30)
if not epaulette:
labels_to_keep.append(31)
if not sleeve:
labels_to_keep.append(32)
if not pocket:
labels_to_keep.append(33)
if not neckline:
labels_to_keep.append(34)
if not buckle:
labels_to_keep.append(35)
if not zipper:
labels_to_keep.append(36)
if not applique:
labels_to_keep.append(37)
if not bead:
labels_to_keep.append(38)
if not bow:
labels_to_keep.append(39)
if not flower:
labels_to_keep.append(40)
if not fringe:
labels_to_keep.append(41)
if not ribbon:
labels_to_keep.append(42)
if not rivet:
labels_to_keep.append(43)
if not ruffle:
labels_to_keep.append(44)
if not sequin:
labels_to_keep.append(45)
if not tassel:
labels_to_keep.append(46)
mask = np.isin(pred_seg, labels_to_keep).astype(np.uint8)
# 创建agnostic-mask图像
mask_image = Image.fromarray(mask * 255)
mask_image = mask_image.convert("RGB")
mask_image = pil2tensor(mask_image)
results.append(mask_image)
return (torch.cat(results, dim=0),)