Update app.py

app.py

@@ -5,21 +5,18 @@ import numpy as np
 import io
 import pandas as pd
 from lime import lime_image
+import time
+
+# Define your image size
+IMG_SIZE = 256  
 
 # Load the model
 def load_model():
     model = tf.keras.models.load_model("custom_model_final.h5", compile=False)
-    # Explicitly compile the model with the required loss function
-    loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False)
-    model.compile(optimizer='adam', loss=loss_fn, metrics=['accuracy'])
+    # Compile the model if necessary
+    # model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])
     return model
 
-# Load the labels
-def load_labels():
-    with open("labels.txt", "r") as file:
-        class_names = [line.strip() for line in file.readlines()]
-    return class_names
-
 # Preprocess image
 def preprocess_image(image):
     image = image.resize((256, 256))
@@ -28,12 +25,18 @@ def preprocess_image(image):
     data = np.expand_dims(normalized_image_array, axis=0)
     return data
 
-# Classify image
-def classify_image(model, image, class_names):
-    prediction = model.predict(image)
-    predicted_class = class_names[np.argmax(prediction[0])]
-    confidence_score = round(100 * np.max(prediction[0]), 2)
-    return predicted_class, confidence_score
+# Define the predict function
+def predict(model, img):
+    img = img.resize((IMG_SIZE, IMG_SIZE))  # Resize the image
+    img_array = tf.keras.preprocessing.image.img_to_array(img)
+    img_array = tf.expand_dims(img_array, 0)
+
+    predictions = model.predict(img_array)
+
+    class_labels = ["normal", "cataract", "retina disease", "glaucoma"]
+    predicted_class = class_labels[np.argmax(predictions[0])]
+    confidence = round(100 * (np.max(predictions[0])), 2)
+    return predicted_class, confidence
 
 # Explain image
 def explain_image(image, model):
@@ -56,15 +59,14 @@ def main():
     st.sidebar.title("Upload Image")
     uploaded_file = st.sidebar.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])
 
-    
     if uploaded_file is not None:
-        # Load model and labels
+        # Load model
         model = load_model()
-        class_names = load_labels()
         
         # Display uploaded image
         image = Image.open(io.BytesIO(uploaded_file.read()))
         st.image(image, caption="Uploaded Image", use_column_width=True)
+        processed_image = preprocess_image(image)
         
         # Predict button
         predict_button = st.sidebar.button("Predict", key="predict_button")
@@ -74,26 +76,26 @@ def main():
             </style>""", unsafe_allow_html=True
         )
         if predict_button:
-            # Preprocess image
-            processed_image = preprocess_image(image)
-            
-            # Classify image
-            predicted_class, confidence_score = classify_image(model, processed_image, class_names)
-            
-            # Explain image classification
-            explanation_image = explain_image(processed_image, model)
-            
-            # Display explanation image
-            st.image(explanation_image, caption="Explanation Image", use_column_width=True)
+            # Display processing message with spinner
+            with st.spinner("Please wait...Processing the image and predicting..."):
+                
+                # Classify image
+                predicted_class, confidence_score = predict(model, image)
 
-            # Display prediction
-            st.subheader("Prediction:")
-            # Create a table for prediction results
-            prediction_table = pd.DataFrame({
-                "Predicted Class": [predicted_class],
-                "Confidence": [f"{confidence_score}%"]
-            })
-            st.table(prediction_table)
+                # Explain image classification
+                explanation_image = explain_image(processed_image, model)
+                
+                # Display explanation image
+                st.image(explanation_image, caption="Explanation Image", use_column_width=True)
+                
+                # Display prediction
+                st.subheader("Prediction:")
+                # Create a table for prediction results
+                prediction_table = pd.DataFrame({
+                    "Predicted Class": [predicted_class],
+                    "Confidence": [f"{confidence_score}%"]
+                })
+                st.table(prediction_table)
 
 if __name__ == "__main__":
-    main()
+    main()