Update app.py

app.py

@@ -6,78 +6,81 @@ from mtcnn import MTCNN
 import cv2
 import io
 
+# Assuming model is included in your repository or loaded from a URL
 # Load the model (ensure correct path for loading)
-# For Hugging Face Spaces, ensure the model file is included in the repository or loaded from a URL
-
-
-input_shape = (224, 224, 3)
-model = tf.keras.models.load_model('oily_dry.h5', compile=False, custom_objects={'input_shape': input_shape})
-
-# model = tf.keras.models.load_model('oily_dry.h5')
+model = tf.keras.models.load_model('oily_dry.h5')
 
 # Load the MTCNN face detection model
 mtcnn = MTCNN()
 
-def detect_and_process_skin(image, output_folder=None):
-    """Detects faces in an image, crops the skin region, and prepares it for classification."""
-    # Convert PIL Image to numpy array
-    img_np = np.array(image)
+def detect_and_process_skin(image_bytes):
+    """Detects faces in an image, crops the skin region, and returns it as an image object."""
+    # Load image from bytes
+    img = Image.open(io.BytesIO(image_bytes))
+    img_np = np.array(img)
     img_rgb = cv2.cvtColor(img_np, cv2.COLOR_BGR2RGB)
 
     # Detect faces in the image
     detections = mtcnn.detect_faces(img_rgb)
 
+    # Check if any faces were detected
     if detections:
         x, y, width, height = detections[0]['box']
 
         # Crop the face region
-        face_img = img_np[y:y+height, x:x+width]
-
-        # Convert cropped face to PIL Image for compatibility with model preprocessing
-        pil_img = Image.fromarray(cv2.cvtColor(face_img, cv2.COLOR_BGR2RGB))
+        face_img_np = img_np[y:y+height, x:x+width]
 
+        # Convert to PIL Image for return
+        pil_img = Image.fromarray(face_img_np)
         return pil_img
     else:
-        # Return the original image if no face was detected
-        return image
+        # Return original image if no face was detected
+        return img
 
-def classify_image(image):
-    """Classifies the image after processing."""
-    processed_image = detect_and_process_skin(image)
+def classify_image(image_bytes):
+    """Loads, preprocess, and classifies the image from bytes."""
+    # Process the image using face detection
+    pil_img = detect_and_process_skin(image_bytes)
 
     # Resize and preprocess the image for the model
-    processed_image = processed_image.resize((224, 224))
-    image_array = tf.keras.preprocessing.image.img_to_array(processed_image) / 255.0
-    image_array = np.expand_dims(image_array, axis=0)
+    image = pil_img.resize((224, 224))
+    image = tf.keras.preprocessing.image.img_to_array(image) / 255.0
+    image = np.expand_dims(image, axis=0)
 
     # Classify the image
-    predictions = model.predict(image_array)
+    predictions = model.predict(image)
+
+    # Process predictions for clarity
     predicted_class = np.argmax(predictions)
     percentages = predictions[0] * 100
+    dry_percentage, normal_percentage, oily_percentage = percentages
 
-    return predicted_class, *percentages
+    return predicted_class, dry_percentage, oily_percentage, normal_percentage
 
 def app():
     st.title("Oily/Dry Skin Level Predictor")
+    st.write("Coded by Manith Jayaba")
+    st.write("This app can measure the oiliness and dryness of your skin")
 
+    # Get the image file
     image_file = st.file_uploader("Upload an image", type=["jpg", "jpeg", "png"])
 
+    # Classify and display the result
     if image_file is not None:
-        # Convert the file to an image
-        image = Image.open(image_file)
-        predicted_class, dry_percentage, oily_percentage, normal_percentage = classify_image(image)
+        predicted_class, dry_percentage, oily_percentage, normal_percentage = classify_image(image_file.getvalue())
 
         # Convert processed image to bytes for display
-        processed_image_bytes = io.BytesIO()
-        image.save(processed_image_bytes, format='JPEG')
-        processed_image_bytes.seek(0)
-
-        st.image(processed_image_bytes, width=250)
+        img_bytes = io.BytesIO()
+        detect_and_process_skin(image_file.getvalue()).save(img_bytes, format='JPEG')
+        st.image(img_bytes.getvalue(), width=250, caption="Processed Image")
 
-        # Display the classification results
+        # Display progress bars
+        st.progress(int(dry_percentage))
         st.write(f"Dry Skin: {dry_percentage:.2f}%")
+        st.progress(int(oily_percentage))
         st.write(f"Oily Skin: {oily_percentage:.2f}%")
+        st.progress(int(normal_percentage))
         st.write(f"Normal Skin: {normal_percentage:.2f}%")
 
-if __name__ == "__main__":
-    app()
+if _name_ == "_main_":
+    app()