Spaces:
Runtime error
Runtime error
hasibzunair
commited on
Commit
ยท
8a3583d
1
Parent(s):
3a4add5
add app files
Browse files- cmap.npy +3 -0
- description.html +10 -0
- nyu.ipynb +165 -0
- requirements.txt +6 -0
- sample_images/a.png +0 -0
- sample_images/b.png +0 -0
- sample_images/c.png +0 -0
- sample_images/d.png +0 -0
- unetplusplus.py +142 -0
cmap.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e5648506a4b5dbeb787e93f26b429cab659c3b66a4d579645edb2f24ba41a919
|
3 |
+
size 848
|
description.html
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
<!DOCTYPE html>
|
2 |
+
<html lang="en">
|
3 |
+
<head>
|
4 |
+
<meta charset="UTF-8">
|
5 |
+
<title>Title</title>
|
6 |
+
</head>
|
7 |
+
<body>
|
8 |
+
This is a demo of our BMVC'2022 Oral paper <a href="https://arxiv.org/abs/2210.00923">Masked Supervised Learning for Semantic Segmentation</a>.</br>
|
9 |
+
</body>
|
10 |
+
</html>
|
nyu.ipynb
ADDED
@@ -0,0 +1,165 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cells": [
|
3 |
+
{
|
4 |
+
"cell_type": "code",
|
5 |
+
"execution_count": null,
|
6 |
+
"metadata": {},
|
7 |
+
"outputs": [],
|
8 |
+
"source": [
|
9 |
+
"import os\n",
|
10 |
+
"import numpy as np\n",
|
11 |
+
"import cv2\n",
|
12 |
+
"import codecs\n",
|
13 |
+
"import torch\n",
|
14 |
+
"import torchvision.transforms as transforms\n",
|
15 |
+
"import gradio as gr\n",
|
16 |
+
"\n",
|
17 |
+
"from PIL import Image\n",
|
18 |
+
"\n",
|
19 |
+
"from unetplusplus import NestedUNet\n",
|
20 |
+
"\n",
|
21 |
+
"torch.manual_seed(0)\n",
|
22 |
+
"\n",
|
23 |
+
"if torch.cuda.is_available():\n",
|
24 |
+
" torch.backends.cudnn.deterministic = True\n",
|
25 |
+
"\n",
|
26 |
+
"# Device\n",
|
27 |
+
"DEVICE = \"cpu\"\n",
|
28 |
+
"print(DEVICE)\n",
|
29 |
+
"\n",
|
30 |
+
"# Load color map\n",
|
31 |
+
"cmap = np.load('cmap.npy')\n",
|
32 |
+
"\n",
|
33 |
+
"# Make directories\n",
|
34 |
+
"os.system(\"mkdir ./models\")\n",
|
35 |
+
"\n",
|
36 |
+
"# Get model weights\n",
|
37 |
+
"if not os.path.exists(\"./models/masksupnyu39.31d.pth\"):\n",
|
38 |
+
" os.system(\"wget -O ./models/masksupnyu39.31d.pth https://github.com/hasibzunair/masksup-segmentation/releases/download/v0.1/masksupnyu39.31iou.pth\")\n",
|
39 |
+
"\n",
|
40 |
+
"# Load model\n",
|
41 |
+
"model = NestedUNet(num_classes=40)\n",
|
42 |
+
"checkpoint = torch.load(\"./models/masksupnyu39.31d.pth\")\n",
|
43 |
+
"model.load_state_dict(checkpoint)\n",
|
44 |
+
"model = model.to(DEVICE)\n",
|
45 |
+
"model.eval()\n",
|
46 |
+
"\n",
|
47 |
+
"\n",
|
48 |
+
"# Main inference function\n",
|
49 |
+
"def inference(img_path):\n",
|
50 |
+
" image = Image.open(img_path).convert(\"RGB\")\n",
|
51 |
+
" transforms_image = transforms.Compose(\n",
|
52 |
+
" [\n",
|
53 |
+
" transforms.Resize((224, 224)),\n",
|
54 |
+
" transforms.CenterCrop((224, 224)),\n",
|
55 |
+
" transforms.ToTensor(),\n",
|
56 |
+
" transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),\n",
|
57 |
+
" ]\n",
|
58 |
+
" )\n",
|
59 |
+
"\n",
|
60 |
+
" image = transforms_image(image)\n",
|
61 |
+
" image = image[None, :]\n",
|
62 |
+
" # Predict\n",
|
63 |
+
" with torch.no_grad():\n",
|
64 |
+
" output = torch.sigmoid(model(image.to(DEVICE).float()))\n",
|
65 |
+
" output = torch.softmax(output, dim=1).argmax(dim=1)[0].float().cpu().numpy().astype(np.uint8)\n",
|
66 |
+
" pred = cmap[output]\n",
|
67 |
+
" return pred\n",
|
68 |
+
"\n",
|
69 |
+
"# App\n",
|
70 |
+
"title = \"Masked Supervised Learning for Semantic Segmentation\"\n",
|
71 |
+
"description = codecs.open(\"description.html\", \"r\", \"utf-8\").read()\n",
|
72 |
+
"article = \"<p style='text-align: center'><a href='https://arxiv.org/abs/2210.00923' target='_blank'>Masked Supervised Learning for Semantic Segmentation</a> | <a href='https://github.com/hasibzunair/masksup-segmentation' target='_blank'>Github</a></p>\"\n",
|
73 |
+
"\n",
|
74 |
+
"gr.Interface(\n",
|
75 |
+
" inference,\n",
|
76 |
+
" gr.inputs.Image(type='file', label=\"Input Image\"),\n",
|
77 |
+
" gr.outputs.Image(type=\"file\", label=\"Predicted Output\"),\n",
|
78 |
+
" examples=[\"./sample_images/a.png\", \"./sample_images/b.png\", \n",
|
79 |
+
" \"./sample_images/c.png\", \"./sample_images/d.png\"],\n",
|
80 |
+
" title=title,\n",
|
81 |
+
" description=description,\n",
|
82 |
+
" article=article,\n",
|
83 |
+
" allow_flagging=False,\n",
|
84 |
+
" analytics_enabled=False,\n",
|
85 |
+
" ).launch(debug=True, enable_queue=True)"
|
86 |
+
]
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"cell_type": "code",
|
90 |
+
"execution_count": null,
|
91 |
+
"metadata": {},
|
92 |
+
"outputs": [],
|
93 |
+
"source": []
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"cell_type": "code",
|
97 |
+
"execution_count": null,
|
98 |
+
"metadata": {},
|
99 |
+
"outputs": [],
|
100 |
+
"source": []
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"cell_type": "code",
|
104 |
+
"execution_count": null,
|
105 |
+
"metadata": {},
|
106 |
+
"outputs": [],
|
107 |
+
"source": []
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"cell_type": "code",
|
111 |
+
"execution_count": null,
|
112 |
+
"metadata": {},
|
113 |
+
"outputs": [],
|
114 |
+
"source": []
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"cell_type": "code",
|
118 |
+
"execution_count": null,
|
119 |
+
"metadata": {},
|
120 |
+
"outputs": [],
|
121 |
+
"source": []
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"cell_type": "code",
|
125 |
+
"execution_count": null,
|
126 |
+
"metadata": {},
|
127 |
+
"outputs": [],
|
128 |
+
"source": []
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"cell_type": "code",
|
132 |
+
"execution_count": null,
|
133 |
+
"metadata": {},
|
134 |
+
"outputs": [],
|
135 |
+
"source": []
|
136 |
+
}
|
137 |
+
],
|
138 |
+
"metadata": {
|
139 |
+
"kernelspec": {
|
140 |
+
"display_name": "Python 3.8.12 ('fifa')",
|
141 |
+
"language": "python",
|
142 |
+
"name": "python3"
|
143 |
+
},
|
144 |
+
"language_info": {
|
145 |
+
"codemirror_mode": {
|
146 |
+
"name": "ipython",
|
147 |
+
"version": 3
|
148 |
+
},
|
149 |
+
"file_extension": ".py",
|
150 |
+
"mimetype": "text/x-python",
|
151 |
+
"name": "python",
|
152 |
+
"nbconvert_exporter": "python",
|
153 |
+
"pygments_lexer": "ipython3",
|
154 |
+
"version": "3.8.12"
|
155 |
+
},
|
156 |
+
"orig_nbformat": 4,
|
157 |
+
"vscode": {
|
158 |
+
"interpreter": {
|
159 |
+
"hash": "5a4cff4f724f20f3784f32e905011239b516be3fadafd59414871df18d0dad63"
|
160 |
+
}
|
161 |
+
}
|
162 |
+
},
|
163 |
+
"nbformat": 4,
|
164 |
+
"nbformat_minor": 2
|
165 |
+
}
|
requirements.txt
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
scipy==1.4.1
|
2 |
+
torch
|
3 |
+
h5py==2.10.0
|
4 |
+
numpy==1.18.1
|
5 |
+
opencv-python-headless==4.2.0.32
|
6 |
+
Pillow
|
sample_images/a.png
ADDED
sample_images/b.png
ADDED
sample_images/c.png
ADDED
sample_images/d.png
ADDED
unetplusplus.py
ADDED
@@ -0,0 +1,142 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch
|
2 |
+
from torch import nn
|
3 |
+
import torch.nn.functional as F
|
4 |
+
|
5 |
+
__all__ = ['UNet', 'NestedUNet']
|
6 |
+
|
7 |
+
"""Taken from https://github.com/4uiiurz1/pytorch-nested-unet"""
|
8 |
+
|
9 |
+
class VGGBlock(nn.Module):
|
10 |
+
def __init__(self, in_channels, middle_channels, out_channels):
|
11 |
+
super().__init__()
|
12 |
+
self.relu = nn.ReLU(inplace=True)
|
13 |
+
self.conv1 = nn.Conv2d(in_channels, middle_channels, 3, padding=1)
|
14 |
+
self.bn1 = nn.BatchNorm2d(middle_channels)
|
15 |
+
self.conv2 = nn.Conv2d(middle_channels, out_channels, 3, padding=1)
|
16 |
+
self.bn2 = nn.BatchNorm2d(out_channels)
|
17 |
+
|
18 |
+
def forward(self, x):
|
19 |
+
out = self.conv1(x)
|
20 |
+
out = self.bn1(out)
|
21 |
+
out = self.relu(out)
|
22 |
+
|
23 |
+
out = self.conv2(out)
|
24 |
+
out = self.bn2(out)
|
25 |
+
out = self.relu(out)
|
26 |
+
|
27 |
+
return out
|
28 |
+
|
29 |
+
|
30 |
+
class UNet(nn.Module):
|
31 |
+
def __init__(self, num_classes, input_channels=3, **kwargs):
|
32 |
+
super().__init__()
|
33 |
+
|
34 |
+
nb_filter = [32, 64, 128, 256, 512]
|
35 |
+
|
36 |
+
self.pool = nn.MaxPool2d(2, 2)
|
37 |
+
self.up = nn.Upsample(scale_factor=2, mode='bilinear', align_corners=True)
|
38 |
+
|
39 |
+
self.conv0_0 = VGGBlock(input_channels, nb_filter[0], nb_filter[0])
|
40 |
+
self.conv1_0 = VGGBlock(nb_filter[0], nb_filter[1], nb_filter[1])
|
41 |
+
self.conv2_0 = VGGBlock(nb_filter[1], nb_filter[2], nb_filter[2])
|
42 |
+
self.conv3_0 = VGGBlock(nb_filter[2], nb_filter[3], nb_filter[3])
|
43 |
+
self.conv4_0 = VGGBlock(nb_filter[3], nb_filter[4], nb_filter[4])
|
44 |
+
|
45 |
+
self.conv3_1 = VGGBlock(nb_filter[3]+nb_filter[4], nb_filter[3], nb_filter[3])
|
46 |
+
self.conv2_2 = VGGBlock(nb_filter[2]+nb_filter[3], nb_filter[2], nb_filter[2])
|
47 |
+
self.conv1_3 = VGGBlock(nb_filter[1]+nb_filter[2], nb_filter[1], nb_filter[1])
|
48 |
+
self.conv0_4 = VGGBlock(nb_filter[0]+nb_filter[1], nb_filter[0], nb_filter[0])
|
49 |
+
|
50 |
+
self.final = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
|
51 |
+
|
52 |
+
|
53 |
+
def forward(self, input):
|
54 |
+
x0_0 = self.conv0_0(input)
|
55 |
+
x1_0 = self.conv1_0(self.pool(x0_0))
|
56 |
+
x2_0 = self.conv2_0(self.pool(x1_0))
|
57 |
+
x3_0 = self.conv3_0(self.pool(x2_0))
|
58 |
+
x4_0 = self.conv4_0(self.pool(x3_0))
|
59 |
+
|
60 |
+
x3_1 = self.conv3_1(torch.cat([x3_0, self.up(x4_0)], 1))
|
61 |
+
x2_2 = self.conv2_2(torch.cat([x2_0, self.up(x3_1)], 1))
|
62 |
+
x1_3 = self.conv1_3(torch.cat([x1_0, self.up(x2_2)], 1))
|
63 |
+
x0_4 = self.conv0_4(torch.cat([x0_0, self.up(x1_3)], 1))
|
64 |
+
|
65 |
+
output = self.final(x0_4)
|
66 |
+
return output
|
67 |
+
|
68 |
+
|
69 |
+
class NestedUNet(nn.Module):
|
70 |
+
"""
|
71 |
+
U-Net Plus plus architecture
|
72 |
+
Reference: https://arxiv.org/abs/1807.10165
|
73 |
+
"""
|
74 |
+
def __init__(self, num_classes=1, input_channels=3, deep_supervision=False, **kwargs):
|
75 |
+
super().__init__()
|
76 |
+
|
77 |
+
nb_filter = [32, 64, 128, 256, 512]
|
78 |
+
|
79 |
+
self.deep_supervision = deep_supervision
|
80 |
+
|
81 |
+
self.pool = nn.MaxPool2d(2, 2)
|
82 |
+
self.up = nn.Upsample(scale_factor=2, mode='bilinear', align_corners=True)
|
83 |
+
|
84 |
+
self.conv0_0 = VGGBlock(input_channels, nb_filter[0], nb_filter[0])
|
85 |
+
self.conv1_0 = VGGBlock(nb_filter[0], nb_filter[1], nb_filter[1])
|
86 |
+
self.conv2_0 = VGGBlock(nb_filter[1], nb_filter[2], nb_filter[2])
|
87 |
+
self.conv3_0 = VGGBlock(nb_filter[2], nb_filter[3], nb_filter[3])
|
88 |
+
self.conv4_0 = VGGBlock(nb_filter[3], nb_filter[4], nb_filter[4])
|
89 |
+
|
90 |
+
self.conv0_1 = VGGBlock(nb_filter[0]+nb_filter[1], nb_filter[0], nb_filter[0])
|
91 |
+
self.conv1_1 = VGGBlock(nb_filter[1]+nb_filter[2], nb_filter[1], nb_filter[1])
|
92 |
+
self.conv2_1 = VGGBlock(nb_filter[2]+nb_filter[3], nb_filter[2], nb_filter[2])
|
93 |
+
self.conv3_1 = VGGBlock(nb_filter[3]+nb_filter[4], nb_filter[3], nb_filter[3])
|
94 |
+
|
95 |
+
self.conv0_2 = VGGBlock(nb_filter[0]*2+nb_filter[1], nb_filter[0], nb_filter[0])
|
96 |
+
self.conv1_2 = VGGBlock(nb_filter[1]*2+nb_filter[2], nb_filter[1], nb_filter[1])
|
97 |
+
self.conv2_2 = VGGBlock(nb_filter[2]*2+nb_filter[3], nb_filter[2], nb_filter[2])
|
98 |
+
|
99 |
+
self.conv0_3 = VGGBlock(nb_filter[0]*3+nb_filter[1], nb_filter[0], nb_filter[0])
|
100 |
+
self.conv1_3 = VGGBlock(nb_filter[1]*3+nb_filter[2], nb_filter[1], nb_filter[1])
|
101 |
+
|
102 |
+
self.conv0_4 = VGGBlock(nb_filter[0]*4+nb_filter[1], nb_filter[0], nb_filter[0])
|
103 |
+
|
104 |
+
if self.deep_supervision:
|
105 |
+
self.final1 = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
|
106 |
+
self.final2 = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
|
107 |
+
self.final3 = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
|
108 |
+
self.final4 = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
|
109 |
+
else:
|
110 |
+
self.final = nn.Conv2d(nb_filter[0], num_classes, kernel_size=1)
|
111 |
+
|
112 |
+
|
113 |
+
def forward(self, input):
|
114 |
+
x0_0 = self.conv0_0(input)
|
115 |
+
x1_0 = self.conv1_0(self.pool(x0_0))
|
116 |
+
x0_1 = self.conv0_1(torch.cat([x0_0, self.up(x1_0)], 1))
|
117 |
+
|
118 |
+
x2_0 = self.conv2_0(self.pool(x1_0))
|
119 |
+
x1_1 = self.conv1_1(torch.cat([x1_0, self.up(x2_0)], 1))
|
120 |
+
x0_2 = self.conv0_2(torch.cat([x0_0, x0_1, self.up(x1_1)], 1))
|
121 |
+
|
122 |
+
x3_0 = self.conv3_0(self.pool(x2_0))
|
123 |
+
x2_1 = self.conv2_1(torch.cat([x2_0, self.up(x3_0)], 1))
|
124 |
+
x1_2 = self.conv1_2(torch.cat([x1_0, x1_1, self.up(x2_1)], 1))
|
125 |
+
x0_3 = self.conv0_3(torch.cat([x0_0, x0_1, x0_2, self.up(x1_2)], 1))
|
126 |
+
|
127 |
+
x4_0 = self.conv4_0(self.pool(x3_0))
|
128 |
+
x3_1 = self.conv3_1(torch.cat([x3_0, self.up(x4_0)], 1))
|
129 |
+
x2_2 = self.conv2_2(torch.cat([x2_0, x2_1, self.up(x3_1)], 1))
|
130 |
+
x1_3 = self.conv1_3(torch.cat([x1_0, x1_1, x1_2, self.up(x2_2)], 1))
|
131 |
+
x0_4 = self.conv0_4(torch.cat([x0_0, x0_1, x0_2, x0_3, self.up(x1_3)], 1))
|
132 |
+
|
133 |
+
if self.deep_supervision:
|
134 |
+
output1 = self.final1(x0_1)
|
135 |
+
output2 = self.final2(x0_2)
|
136 |
+
output3 = self.final3(x0_3)
|
137 |
+
output4 = self.final4(x0_4)
|
138 |
+
return [output1, output2, output3, output4]
|
139 |
+
|
140 |
+
else:
|
141 |
+
output = self.final(x0_4)
|
142 |
+
return output
|