Update app.py

app.py

@@ -78,16 +78,6 @@ cnn.fit(XCnn_train, y_train)
 # Set the page title
 st.title("Handwritten Text Digit Recognition")
 
-# Create a CSS style to center-align text
-centered_title = f"""
-    <div style="text-align: center; padding: 10px;">
-        <h1>Handwritten Text Digit Recognition</h1>
-    </div>
-"""
-# Display the centered title
-st.markdown(centered_title, unsafe_allow_html=True)
-
-
 stroke_width = st.sidebar.slider("Stroke width: ", 1, 35, 32)
 stroke_color = st.sidebar.color_picker("Stroke color hex: ")
 bg_color = st.sidebar.color_picker("Background color hex: ", "#eee")
@@ -111,28 +101,16 @@ canvas_result = st_canvas(
 
 # Do something interesting with the image data and paths
 if canvas_result.image_data is not None:
-    # Preprocess the drawn image
     image = canvas_result.image_data
     image1 = image.copy()
     image1 = image1.astype('uint8')
-
-    # Assuming the background is white, you can invert the colors if necessary
-    image1 = 255 - image1
-
-    # Resize the image to match the input size of your CNN model (28x28)
+    image1 = cv2.cvtColor(image1, cv2.COLOR_BGR2GRAY)
     image1 = cv2.resize(image1, (28, 28))
-    
-    # Normalize the image to values between 0 and 1
-    image1 = image1 / 255.0
-    
-    # Convert the image to the expected shape for the model
-    image1 = image1.reshape(1, 1, 28, 28).astype('float32')
+    st.image(image1)
 
-    # Pass the preprocessed image to your model for prediction
-    prediction = np.argmax(cnn.predict(image1))
-    
-    # Display the prediction result
-    st.title(f"Predicted Digit: {prediction}")
+    # Correctly reshape the image
+    image1 = image1.reshape(1, 1, 28, 28).astype('float32')
+    st.title(np.argmax(model.predict(image1)))
 
 if canvas_result.json_data is not None:
     st.dataframe(pd.json_normalize(canvas_result.json_data["objects"]))