Update app.py

app.py

@@ -1,44 +1,85 @@
 import gradio as gr
-from transformers import pipeline
+from transformers import pipeline 
 import numpy as np
 from PIL import Image
 import tensorflow as tf
 
-model_path = r'./model12_acc99_kera.h5'
+# model_path = r'./model12_acc99_kera.h5'
 
-# Function to preprocess the image and make predictions
-def classify_lung_and_colon(input_image):
-    try:
-        # Preprocess the image (resize, normalize, etc.)
-        input_image = np.array(input_image)
-        input_image_copy = input_image.copy()  # Making a copy to avoid the array reference issue
-        input_image_resized = np.array(Image.fromarray(input_image_copy).resize((228,228))) 
-        # / 255.0
-        input_image_resized = np.expand_dims(input_image_resized, axis=0)
+# # Function to preprocess the image and make predictions
+# def classify_lung_and_colon(input_image):
+#     try:
+#         # Preprocess the image (resize, normalize, etc.)
+#         input_image = np.array(input_image)
+#         input_image_copy = input_image.copy()  # Making a copy to avoid the array reference issue
+#         input_image_resized = np.array(Image.fromarray(input_image_copy).resize((228,228))) 
+#         # / 255.0
+#         input_image_resized = np.expand_dims(input_image_resized, axis=0)
 
-        # Making predictions
-        model_h5 = tf.keras.models.load_model(model_path)
-        predictions = model_h5.predict(input_image_resized)
+#         # Making predictions
+#         model_h5 = tf.keras.models.load_model(model_path)
+#         predictions = model_h5.predict(input_image_resized)
 
-        # Getting the class with the highest probability
-        class_idx = np.argmax(predictions)
-        class_label = ["Colon Adenocarcinoma", "Colon Benign Tissue", "Lung Adenocarcinoma", "Lung Benign Tissue","Lung Squamous Cell Carcinoma"][class_idx]
-        confidence = predictions[0][class_idx]
+#         # Getting the class with the highest probability
+#         class_idx = np.argmax(predictions)
+#         class_label = ["Colon Adenocarcinoma", "Colon Benign Tissue", "Lung Adenocarcinoma", "Lung Benign Tissue","Lung Squamous Cell Carcinoma"][class_idx]
+#         confidence = predictions[0][class_idx]
+
+#         return f"Prediction: {class_label}, Confidence: {confidence:.2f}"
+#     except Exception as e:
+#         return str(e)
+
+# # Creating a Gradio interface
+# iface = gr.Interface(
+#     fn=classify_lung_and_colon,
+#     inputs="image",
+#     outputs="text",
+#     title="Lung and Colon Cancer Detection",
+#     description="Upload a Histopathology Image",
+#     flagging_options = ["Wrong Prediction"],
+#     theme = 'darkhuggingface'
+# )
+
+# # Launching the Gradio interface
+# iface.launch(inline = False)
+import tensorflow as tf
+import numpy as np
+from PIL import Image
+
+# Load the trained model
+model = tf.keras.models.load_model("./model12_acc99_kera.h5")
+
+# Define image preprocessing function
+def preprocess_image(image):
+    # Resize the image to match the input size of the model
+    img = image.resize((224, 224))
+    # Convert image to numpy array
+    img_array = np.array(img)
+    # Normalize pixel values to the range [0, 1]
+    img_array = img_array / 255.0
+    # Expand dimensions to match the model's expected input shape
+    img_array = np.expand_dims(img_array, axis=0)
+    return img_array
+
+# Define the prediction function
+def predict(image):
+    # Preprocess the image
+    processed_image = preprocess_image(image)
+    # Perform inference
+    predictions = model.predict(processed_image)
+    # Convert predictions to class labels
+    class_labels = ['Colon Adenocarcinoma', 'Colon Benign Tissue', 'Lung Adenocarcinoma', 'Lung Benign Tissue', 'Lung Squamous Cell Carcinoma']
+    predicted_class = class_labels[np.argmax(predictions)]
+    return predicted_class
 
-        return f"Prediction: {class_label}, Confidence: {confidence:.2f}"
-    except Exception as e:
-        return str(e)
 
 # Creating a Gradio interface
 iface = gr.Interface(
-    fn=classify_lung_and_colon,
+    fn=predict,
     inputs="image",
     outputs="text",
     title="Lung and Colon Cancer Detection",
     description="Upload a Histopathology Image",
     flagging_options = ["Wrong Prediction"],
-    theme = 'darkhuggingface'
-)
-
-# Launching the Gradio interface
-iface.launch(inline = False)
+    # theme = 'darkhuggingface'
+)