Update app.py
Browse files
app.py
CHANGED
@@ -1,12 +1,14 @@
|
|
1 |
import cv2
|
|
|
2 |
import numpy as np
|
3 |
import torch
|
4 |
-
from fastapi import FastAPI, UploadFile, File
|
5 |
from PIL import Image
|
|
|
|
|
|
|
6 |
from TranSalNet_Res import TranSalNet
|
7 |
from utils.data_process import preprocess_img, postprocess_img
|
8 |
|
9 |
-
|
10 |
app = FastAPI()
|
11 |
|
12 |
device = torch.device('cpu')
|
@@ -15,51 +17,50 @@ model.load_state_dict(torch.load('pretrained_models/TranSalNet_Res.pth', map_loc
|
|
15 |
model.to(device)
|
16 |
model.eval()
|
17 |
|
|
|
18 |
def count_and_label_red_patches(heatmap, threshold=200):
|
19 |
red_mask = heatmap[:, :, 2] > threshold
|
20 |
contours, _ = cv2.findContours(red_mask.astype(np.uint8), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
|
21 |
-
|
22 |
-
# Sort the contours based on their areas in descending order
|
23 |
contours = sorted(contours, key=cv2.contourArea, reverse=True)
|
24 |
-
|
25 |
original_image = np.array(image)
|
26 |
-
|
27 |
-
centroid_list = []
|
28 |
-
|
29 |
for i, contour in enumerate(contours, start=1):
|
30 |
-
# Compute the centroid of the current contour
|
31 |
M = cv2.moments(contour)
|
32 |
if M["m00"] != 0:
|
33 |
cX = int(M["m10"] / M["m00"])
|
34 |
cY = int(M["m01"] / M["m00"])
|
35 |
else:
|
36 |
cX, cY = 0, 0
|
37 |
-
|
38 |
-
radius = 20
|
39 |
-
circle_color = (0, 0, 0)
|
40 |
-
cv2.circle(original_image, (cX, cY), radius, circle_color, -1)
|
41 |
|
42 |
font = cv2.FONT_HERSHEY_SIMPLEX
|
43 |
font_scale = 1
|
44 |
font_color = (255, 255, 255)
|
45 |
line_type = cv2.LINE_AA
|
46 |
cv2.putText(original_image, str(i), (cX - 10, cY + 10), font, font_scale, font_color, 2, line_type)
|
47 |
-
|
48 |
-
centroid_list.append((cX, cY)) # Add the centroid to the list
|
49 |
|
50 |
-
|
|
|
51 |
for i in range(len(centroid_list) - 1):
|
52 |
start_point = centroid_list[i]
|
53 |
end_point = centroid_list[i + 1]
|
54 |
-
line_color = (0, 0, 0)
|
55 |
cv2.line(original_image, start_point, end_point, line_color, 2)
|
56 |
|
57 |
return original_image, len(contours)
|
58 |
|
|
|
59 |
def process_image(image: Image.Image) -> np.ndarray:
|
60 |
img = image.resize((384, 288))
|
61 |
img = np.array(img)
|
62 |
-
img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
|
63 |
img = np.array(img) / 255.
|
64 |
img = np.expand_dims(np.transpose(img, (2, 0, 1)), axis=0)
|
65 |
img = torch.from_numpy(img)
|
@@ -68,7 +69,7 @@ def process_image(image: Image.Image) -> np.ndarray:
|
|
68 |
pred_saliency = model(img).squeeze().detach().numpy()
|
69 |
|
70 |
heatmap = (pred_saliency * 255).astype(np.uint8)
|
71 |
-
heatmap = cv2.applyColorMap(heatmap, cv2.COLORMAP_JET)
|
72 |
|
73 |
heatmap = cv2.resize(heatmap, (image.width, image.height))
|
74 |
|
@@ -88,17 +89,17 @@ def process_image(image: Image.Image) -> np.ndarray:
|
|
88 |
|
89 |
return blended_img
|
90 |
|
91 |
-
@app.post("/process_image")
|
92 |
-
async def process_uploaded_image(file: UploadFile = File(...)):
|
93 |
-
try:
|
94 |
-
contents = await file.read()
|
95 |
-
image = Image.open(io.BytesIO(contents))
|
96 |
-
except Exception as e:
|
97 |
-
raise HTTPException(status_code=400, detail=f"Error opening image: {str(e)}")
|
98 |
|
99 |
-
|
100 |
-
|
101 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
102 |
|
103 |
-
|
104 |
-
raise HTTPException(status_code=500, detail=f"Error processing image: {str(e)}")
|
|
|
1 |
import cv2
|
2 |
+
import io
|
3 |
import numpy as np
|
4 |
import torch
|
|
|
5 |
from PIL import Image
|
6 |
+
from fastapi import FastAPI, UploadFile, File, HTTPException
|
7 |
+
from gradio import Gradio, Image as GImage
|
8 |
+
from starlette.responses import StreamingResponse
|
9 |
from TranSalNet_Res import TranSalNet
|
10 |
from utils.data_process import preprocess_img, postprocess_img
|
11 |
|
|
|
12 |
app = FastAPI()
|
13 |
|
14 |
device = torch.device('cpu')
|
|
|
17 |
model.to(device)
|
18 |
model.eval()
|
19 |
|
20 |
+
|
21 |
def count_and_label_red_patches(heatmap, threshold=200):
|
22 |
red_mask = heatmap[:, :, 2] > threshold
|
23 |
contours, _ = cv2.findContours(red_mask.astype(np.uint8), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
|
24 |
+
|
|
|
25 |
contours = sorted(contours, key=cv2.contourArea, reverse=True)
|
26 |
+
|
27 |
original_image = np.array(image)
|
28 |
+
|
29 |
+
centroid_list = []
|
30 |
+
|
31 |
for i, contour in enumerate(contours, start=1):
|
|
|
32 |
M = cv2.moments(contour)
|
33 |
if M["m00"] != 0:
|
34 |
cX = int(M["m10"] / M["m00"])
|
35 |
cY = int(M["m01"] / M["m00"])
|
36 |
else:
|
37 |
cX, cY = 0, 0
|
38 |
+
|
39 |
+
radius = 20
|
40 |
+
circle_color = (0, 0, 0)
|
41 |
+
cv2.circle(original_image, (cX, cY), radius, circle_color, -1)
|
42 |
|
43 |
font = cv2.FONT_HERSHEY_SIMPLEX
|
44 |
font_scale = 1
|
45 |
font_color = (255, 255, 255)
|
46 |
line_type = cv2.LINE_AA
|
47 |
cv2.putText(original_image, str(i), (cX - 10, cY + 10), font, font_scale, font_color, 2, line_type)
|
|
|
|
|
48 |
|
49 |
+
centroid_list.append((cX, cY))
|
50 |
+
|
51 |
for i in range(len(centroid_list) - 1):
|
52 |
start_point = centroid_list[i]
|
53 |
end_point = centroid_list[i + 1]
|
54 |
+
line_color = (0, 0, 0)
|
55 |
cv2.line(original_image, start_point, end_point, line_color, 2)
|
56 |
|
57 |
return original_image, len(contours)
|
58 |
|
59 |
+
|
60 |
def process_image(image: Image.Image) -> np.ndarray:
|
61 |
img = image.resize((384, 288))
|
62 |
img = np.array(img)
|
63 |
+
img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
|
64 |
img = np.array(img) / 255.
|
65 |
img = np.expand_dims(np.transpose(img, (2, 0, 1)), axis=0)
|
66 |
img = torch.from_numpy(img)
|
|
|
69 |
pred_saliency = model(img).squeeze().detach().numpy()
|
70 |
|
71 |
heatmap = (pred_saliency * 255).astype(np.uint8)
|
72 |
+
heatmap = cv2.applyColorMap(heatmap, cv2.COLORMAP_JET)
|
73 |
|
74 |
heatmap = cv2.resize(heatmap, (image.width, image.height))
|
75 |
|
|
|
89 |
|
90 |
return blended_img
|
91 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
92 |
|
93 |
+
def gr_process_image(input_image):
|
94 |
+
image = Image.fromarray(input_image)
|
95 |
+
processed_image = process_image(image)
|
96 |
+
return processed_image
|
97 |
+
|
98 |
+
|
99 |
+
iface = Gradio.Interface(
|
100 |
+
fn=gr_process_image,
|
101 |
+
inputs=GImage(),
|
102 |
+
outputs=GImage("numpy")
|
103 |
+
)
|
104 |
|
105 |
+
iface.launch(share=True)
|
|