Update app.py

app.py

@@ -5,23 +5,32 @@ import numpy as np
 from PIL import Image
 
 # Load the models
-filter_model = load_model("model_checkpoint")  # Replace with actual directory path
-corn_model = load_model("corn_model_mobilenetv2.h5")
-cassava_model = load_model("cassava_disease_model.h5")
+filter_model = load_model("model_checkpoint")
+corn_model = load_model('corn_model_mobilenetv2.h5')
+cassava_model = load_model('cassava_disease_model.h5')
 
 # Function to preprocess and predict with the filter model
 def is_leaf_image(img, model, threshold=0.5):
     img_resized = img.resize((128, 128))  # Ensure correct size for the filter model
     img_array = np.expand_dims(image.img_to_array(img_resized) / 255.0, axis=0)  # Normalize
     prediction = model.predict(img_array)
-    leaf_probability = prediction[0][1]  # Assuming "Leaf" class is at index 1
+    
+    # Debugging output
+    st.write(f"Filter Model Prediction Output: {prediction}")
+    
+    # Handle single-value or two-class outputs
+    if len(prediction[0]) == 1:  # Single-class output
+        leaf_probability = prediction[0][0]  # Output is the probability of "Leaf"
+    else:  # Two-class output
+        leaf_probability = prediction[0][1]  # "Leaf" class probability
+
     return leaf_probability >= threshold, leaf_probability
 
-# Function to predict crop disease
-def predict_disease(img, model, class_names):
+# Function to predict an uploaded image
+def predict_image(img, model, class_names):
     img = image.img_to_array(img)
     img = np.expand_dims(img, axis=0)
-    img = img / 255.0  # Normalize
+    img = img / 255.0  # Normalize image
     prediction = model.predict(img)
     predicted_class = np.argmax(prediction)
     return predicted_class, class_names[predicted_class]
@@ -33,6 +42,11 @@ def home_page():
     st.write("""
         Welcome to the Crop Disease Identification App. This tool aims at helping local farmers
         identify diseases in maize and cassava crops by analyzing images of leaves.
+
+        Use the navigation menu to explore the app:
+        - **Home:** Overview of the app
+        - **About:** Information about the app and its purpose
+        - **Prediction:** Upload an image and get a disease diagnosis along with recommendations
     """)
 
 def about_page():
@@ -40,57 +54,68 @@ def about_page():
     st.image("IMG_20240727_143208_444.jpg", use_column_width=True)
     st.write("""
         This application is designed to assist farmers in diagnosing common diseases in maize and cassava crops.
-        The models were built using Convolutional Neural Networks (CNNs), trained on publicly available datasets.
+        By utilizing deep learning models trained on thousands of leaf images, the app can accurately identify
+        diseases such as "blight, common rust and gray leaf spot" in maize and "Bacterial blight, green mottle, brown streak and mosaic" in cassava.
+        The models were built using Convolutional Neural Networks (CNNs), trained on publicly available datasets and tested on locally obtained images.
+        The goal is to provide quick and accurate disease identification to help mitigate crop losses and improve yields.
+
+        If you have any questions or feedback, please reach out to the developer.
+
+        E-mail: olorunnisholato7@gmail.com
+
+        Thank you for using the Crop Disease Identification App!
     """)
 
 def prediction_page():
     st.title("Crop Disease Prediction")
 
-    # Select crop type
+    # Model selection
     crop_type = st.selectbox("Select the crop:", ["Maize", "Cassava"])
 
-    # File uploader
+    # File uploader for image input
     uploaded_file = st.file_uploader("Upload an image of the leaf...", type=["jpg", "jpeg", "png"])
 
     if uploaded_file is not None:
-        # Load and display the uploaded image
+        # Load the uploaded image
         img = Image.open(uploaded_file)
-        st.image(img, caption="Uploaded Image", use_column_width=True)
-
-        if st.button("Classify"):
-            st.write("Classifying...")
-
-            # Filter model prediction
-            is_leaf, leaf_probability = is_leaf_image(img, filter_model)
-            if not is_leaf:
-                st.warning(f"The uploaded image does not appear to contain a leaf. (Confidence: {leaf_probability:.2f})")
-                return
-
-            # Crop-specific model prediction
+        st.image(img, caption='Uploaded Image', use_column_width=True)
+        st.write("")
+
+        # Check if the image contains a leaf
+        is_leaf, leaf_probability = is_leaf_image(img, filter_model)
+
+        if not is_leaf:
+            st.error(f"This image does not appear to contain a leaf (Probability: {leaf_probability:.2f}).")
+        else:
+            st.success(f"The image contains a leaf (Probability: {leaf_probability:.2f}).")
+            
+            # Set image size based on crop type
             if crop_type == "Maize":
-                img_resized = img.resize((128, 128))
-                class_names = ["Blight", "Common_Rust", "Gray_Leaf_Spot", "Healthy"]
-                _, class_name = predict_disease(img_resized, corn_model, class_names)
+                img = img.resize((128, 128))  # Resize for Maize
+                class_names = ['Blight', 'Common_Rust', 'Gray_Leaf_Spot', 'Healthy']
+                predicted_class, class_name = predict_image(img, corn_model, class_names)
             else:
-                img_resized = img.resize((224, 224))
-                class_names = ["Cassava_bacterial_blight", "Cassava_brown_streak_disease", "Cassava_green_mottle", "Cassava_healthy", "Cassava_mosaic_disease"]
-                _, class_name = predict_disease(img_resized, cassava_model, class_names)
+                img = img.resize((224, 224))  # Resize for Cassava
+                class_names = ['Cassava_bacterial_blight','Cassava_brown_streak_disease','Cassava_green_mottle','Cassava_healthy','Cassava_mosaic_disease']
+                predicted_class, class_name = predict_image(img, cassava_model, class_names)
 
-            # Display result
+            # Display prediction
             st.write(f"Predicted Class: {class_name}")
-            if "Healthy" in class_name:
-                st.success("The leaf is healthy. No action needed.")
-            else:
-                recommendations = {
-                    "Blight": "Recommendation: Remove infected plants and apply fungicides.",
-                    "Common_Rust": "Recommendation: Use resistant varieties and apply fungicides if necessary.",
-                    "Gray_Leaf_Spot": "Recommendation: Ensure proper crop rotation and apply fungicides.",
-                    "Cassava_bacterial_blight": "Recommendation: Remove infected plants and use disease-free planting material.",
-                    "Cassava_brown_streak_disease": "Recommendation: Plant resistant cassava varieties.",
-                    "Cassava_green_mottle": "Recommendation: Ensure good field sanitation.",
-                    "Cassava_mosaic_disease": "Recommendation: Use resistant varieties and practice good hygiene."
-                }
-                st.warning(recommendations.get(class_name, "No specific recommendation available."))
+
+            # Recommendations based on prediction
+            recommendations = {
+                'Blight': "Remove infected plants, rotate crops, apply fungicides.",
+                'Common_Rust': "Use resistant varieties, apply fungicides early.",
+                'Gray_Leaf_Spot': "Monitor fields, apply fungicides, manage nitrogen.",
+                'Cassava_bacterial_blight': "Remove infected plants, sterilize tools.",
+                'Cassava_brown_streak_disease': "Plant resistant varieties, manage whiteflies.",
+                'Cassava_green_mottle': "Use resistant varieties, control weeds.",
+                'Cassava_mosaic_disease': "Plant resistant varieties, ensure hygiene.",
+                'Healthy': "The leaf is healthy. No action needed.",
+                'Cassava_healthy': "The leaf is healthy. No action needed."
+            }
+            if class_name in recommendations:
+                st.warning(recommendations[class_name])
 
 # Main app
 def main():