Spaces:
Sleeping
Sleeping
Commit
·
18b3012
1
Parent(s):
ceb4a0e
Update app.py
Browse files
app.py
CHANGED
@@ -1,68 +1,68 @@
|
|
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 |
-
|
|
|
1 |
+
import cv2
|
2 |
+
import sys
|
3 |
+
import json
|
4 |
+
import torch
|
5 |
+
import warnings
|
6 |
+
import numpy as np
|
7 |
+
import streamlit as st
|
8 |
+
import matplotlib.pyplot as plt
|
9 |
+
from segment_anything import sam_model_registry, SamAutomaticMaskGenerator, SamPredictor
|
10 |
+
warnings.filterwarnings('ignore')
|
11 |
|
12 |
+
@st.cache_data()
|
13 |
+
def mask_generate():
|
14 |
+
'''
|
15 |
+
Generate mask for image segmentation
|
16 |
+
'''
|
17 |
+
sam_checkpoint = "assets\model\sam_vit_l_0b3195.pth"
|
18 |
+
model_type = "vit_l"
|
19 |
+
device = "cpu"
|
20 |
|
21 |
+
sam = sam_model_registry[model_type](checkpoint=sam_checkpoint)
|
22 |
+
sam.to(device=device)
|
23 |
+
mask_generator = SamAutomaticMaskGenerator(sam)
|
24 |
+
return mask_generator
|
25 |
|
26 |
|
27 |
+
def show_annot(annot, ax):
|
28 |
+
'''
|
29 |
+
Show annotations on image
|
30 |
+
'''
|
31 |
+
if len(annot) == 0:
|
32 |
+
return
|
33 |
+
sorted_annot = sorted(annot, key=(lambda x: x['area']), reverse=True)
|
34 |
+
polygons = []
|
35 |
+
color = []
|
36 |
+
for ann in sorted_annot:
|
37 |
+
m = ann['segmentation']
|
38 |
+
img = np.ones((m.shape[0], m.shape[1], 3))
|
39 |
+
color_mask = np.random.random((1, 3)).tolist()[0]
|
40 |
+
for i in range(3):
|
41 |
+
img[:,:,i] = color_mask[i]
|
42 |
+
ax.imshow(np.dstack((img, m*0.35)))
|
43 |
|
44 |
+
print(torch.cuda.is_available())
|
45 |
|
46 |
+
st.title("Segment Anything Model (SAM)")
|
47 |
+
image_path = st.file_uploader("Upload Image")
|
48 |
+
if image_path:
|
49 |
+
with st.spinner("Segmenting image..."):
|
50 |
+
image = cv2.imdecode(np.fromstring(image_path.read(), np.uint8), 1)
|
51 |
+
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
|
52 |
|
53 |
+
mask_generator = mask_generate()
|
54 |
+
masks = mask_generator.generate(image)
|
55 |
|
56 |
+
col_original, col_annot = st.columns(2)
|
57 |
+
with col_original:
|
58 |
+
st.image(image)
|
59 |
+
st.caption("Original Image")
|
60 |
+
with col_annot:
|
61 |
+
fig, ax = plt.subplots(figsize=(20,20))
|
62 |
+
ax.imshow(image)
|
63 |
+
show_annot(masks, ax)
|
64 |
+
ax.axis('off')
|
65 |
+
st.pyplot(fig)
|
66 |
+
st.caption("Output Image")
|
67 |
+
else:
|
68 |
+
st.warning('Upload an Image')
|