Update app.py
Browse files
app.py
CHANGED
@@ -3,12 +3,6 @@ from tensorflow.keras.models import load_model
|
|
3 |
import cv2
|
4 |
import numpy as np
|
5 |
|
6 |
-
# Load the trained model
|
7 |
-
model = load_model('FightOS_CNN_Models.h5')
|
8 |
-
|
9 |
-
# Define the class labels
|
10 |
-
class_labels = ['positive', 'negative']
|
11 |
-
|
12 |
# Function to preprocess the image (resize, normalize, etc.)
|
13 |
def preprocess_image(image):
|
14 |
# Resize the image to 224x224
|
@@ -22,7 +16,7 @@ def preprocess_image(image):
|
|
22 |
|
23 |
return preprocessed_image
|
24 |
|
25 |
-
# Function to segment the image using Otsu's thresholding
|
26 |
def segment_image(image):
|
27 |
# Convert the image to grayscale
|
28 |
grayscale_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
|
@@ -30,31 +24,38 @@ def segment_image(image):
|
|
30 |
# Apply Otsu's thresholding
|
31 |
_, thresholded_image = cv2.threshold(grayscale_image, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)
|
32 |
|
33 |
-
|
|
|
|
|
|
|
|
|
|
|
34 |
|
35 |
# Function to make the classification prediction and return the segmented image
|
36 |
def classify_pcos(image):
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
|
|
|
|
|
58 |
|
59 |
# Gradio app interface with multiple outputs
|
60 |
iface = gr.Interface(
|
|
|
3 |
import cv2
|
4 |
import numpy as np
|
5 |
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
# Function to preprocess the image (resize, normalize, etc.)
|
7 |
def preprocess_image(image):
|
8 |
# Resize the image to 224x224
|
|
|
16 |
|
17 |
return preprocessed_image
|
18 |
|
19 |
+
# Function to segment the image using Otsu's thresholding and post-processing
|
20 |
def segment_image(image):
|
21 |
# Convert the image to grayscale
|
22 |
grayscale_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
|
|
|
24 |
# Apply Otsu's thresholding
|
25 |
_, thresholded_image = cv2.threshold(grayscale_image, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)
|
26 |
|
27 |
+
# Perform morphological operations for post-processing
|
28 |
+
kernel = np.ones((5, 5), np.uint8)
|
29 |
+
segmented_image = cv2.morphologyEx(thresholded_image, cv2.MORPH_OPEN, kernel, iterations=2)
|
30 |
+
segmented_image = cv2.morphologyEx(segmented_image, cv2.MORPH_CLOSE, kernel, iterations=2)
|
31 |
+
|
32 |
+
return segmented_image
|
33 |
|
34 |
# Function to make the classification prediction and return the segmented image
|
35 |
def classify_pcos(image):
|
36 |
+
# Preprocess the image (resize, normalize, etc.)
|
37 |
+
preprocessed_image = preprocess_image(image)
|
38 |
+
|
39 |
+
# Make the prediction using the model
|
40 |
+
prediction = model.predict(preprocessed_image)
|
41 |
+
|
42 |
+
# Get the predicted class
|
43 |
+
class_index = np.argmax(prediction)
|
44 |
+
class_label = class_labels[class_index]
|
45 |
+
|
46 |
+
# Get the probability of each class
|
47 |
+
probabilities = {label: round(float(prediction[0][i]), 2) for i, label in enumerate(class_labels)}
|
48 |
+
|
49 |
+
# Segment the image
|
50 |
+
segmented_image = segment_image(image)
|
51 |
+
|
52 |
+
return class_label, probabilities, segmented_image
|
53 |
+
|
54 |
+
# Load the trained model
|
55 |
+
model = load_model('FightOS_CNN_Models.h5')
|
56 |
+
|
57 |
+
# Define the class labels
|
58 |
+
class_labels = ['positive', 'negative']
|
59 |
|
60 |
# Gradio app interface with multiple outputs
|
61 |
iface = gr.Interface(
|