Enhance image preprocessing documentation and clarify 3D model generation process. Update function docstrings to include detailed descriptions of input, output, and processing steps.

app.py

@@ -36,28 +36,32 @@ pipeline.preprocess_image(Image.fromarray(np.zeros((512, 512, 3), dtype=np.uint8
 
 
 def preprocess_image(image: Image.Image) -> Image.Image:
-    """Preprocess the input image.
+    """Preprocess the input image for 3D model generation.
 
-    Args:
-        image (Image.Image): The input image.
-
-    Returns:
-        Image.Image: The preprocessed image.
-    """
-    return pipeline.preprocess_image(image)
-
-
-def preprocess_images(images: list[tuple[Image.Image, str]]) -> list[Image.Image]:
-    """Preprocess a list of input images.
+    This function performs several preprocessing steps to prepare the image for 3D model generation:
+    1. Handles alpha channel or removes background if not present
+    2. Centers and crops the object
+    3. Normalizes the image size to 518x518 pixels
+    4. Applies proper alpha channel processing
 
     Args:
-        images (List[Tuple[Image.Image, str]]): The input images.
+        image (Image.Image): The input image to be preprocessed. Can be either RGB or RGBA format.
 
     Returns:
-        List[Image.Image]: The preprocessed images.
+        Image.Image: The preprocessed image with the following characteristics:
+            - Size: 518x518 pixels
+            - Format: RGBA
+            - Background: Removed
+            - Object: Centered and properly scaled
+
+    Raises:
+        None: This function does not raise any exceptions.
+
+    Note:
+        The preprocessing is handled by the pipeline's internal preprocessing function,
+        which uses rembg for background removal if needed.
     """
-    images = [image[0] for image in images]
-    return [pipeline.preprocess_image(image) for image in images]
+    return pipeline.preprocess_image(image)
 
 
 def save_state_to_file(gs: Gaussian, mesh: MeshExtractResult, output_path: str) -> None:
@@ -118,6 +122,10 @@ def image_to_3d(
 ) -> tuple[str, str]:
     """Convert an image to a 3D model.
 
+    This function takes an input image and generates a 3D model using a two-stage process
+    with separate parameters for each stage. It also generates a preview video that combines
+    color and normal map renderings of the 3D model.
+
     Args:
         image (Image.Image): The input image.
         seed (int): The random seed.
@@ -127,8 +135,13 @@ def image_to_3d(
         slat_sampling_steps (int): The number of sampling steps for structured latent generation.
 
     Returns:
-        str: The path to the pickle file that contains the state of the generated 3D model.
-        str: The path to the video of the 3D model.
+        tuple[str, str]: A tuple containing:
+            - str: Path to the state file (.pth) containing the 3D model data
+            - str: Path to the preview video file (.mp4) showing the 3D model rotation
+
+    Note:
+        The generated files are saved as temporary files that will not be automatically
+        deleted. It is the caller's responsibility to manage these files.
     """
     outputs = pipeline.run(
         image,