Update app.py

app.py

@@ -12,36 +12,30 @@ model = tf.keras.models.load_model(model_path)
 
 # Step 3: Function to preprocess the input image
 def load_and_preprocess_image(img, target_size=(256, 256)):
-    # Resize the image to the model's expected input size
     img = img.resize(target_size)
-    
-    # Convert to array and normalize
     img_array = np.array(img) / 255.0
-    
-    # Expand dimensions to match the input shape of the model
     img_array = np.expand_dims(img_array, axis=0)
-    
     return img_array
 
 # Step 4: Function to make predictions
 def predict_image(img):
-    # Preprocess the image
     img_array = load_and_preprocess_image(img)
-    
-    # Make a prediction
-    prediction = model.predict(img_array)[0][0]
-    
-    # Confidence scores
-    real_confidence = prediction * 100
-    fake_confidence = (1 - prediction) * 100
-    
-    # Determine label
+    prediction = model.predict(img_array)
+
+    # Determine output shape and extract confidence
+    if prediction.shape[-1] == 1:
+        prob_real = prediction[0][0]
+    elif prediction.shape[-1] == 2:
+        prob_real = prediction[0][1]  # Assuming second index = "real"
+    else:
+        return "Model output shape not recognized.", "Cannot compute prediction."
+
+    real_confidence = prob_real * 100
+    fake_confidence = (1 - prob_real) * 100
     result_label = "Real" if real_confidence > fake_confidence else "Fake"
-    
-    # Return results
+
     result_text = f"The model predicts this image is '{result_label}' with {max(real_confidence, fake_confidence):.2f}% confidence."
     explanation = f"Real Confidence: {real_confidence:.2f}% | Fake Confidence: {fake_confidence:.2f}%"
-    
     return result_text, explanation
 
 # Step 5: Define the Gradio interface
@@ -49,7 +43,8 @@ interface = gr.Interface(
     fn=predict_image,
     inputs=gr.Image(type="pil", label="Upload an Image"),
     outputs=[
-        gr.Textbox(label="Prediction Result")
+        gr.Textbox(label="Prediction Result"),
+        gr.Textbox(label="Confidence Scores")
     ],
     title="Deepfake Image Detector",
     description="Upload an image, and the model will classify whether it is a 'real' or 'fake' image using deep learning."
@@ -58,3 +53,4 @@ interface = gr.Interface(
 # Step 6: Launch the app
 if __name__ == "__main__":
     interface.launch()
+