add windowing

app.py

@@ -155,6 +155,40 @@ def to_numpy_image(image: Any) -> np.ndarray:
     return arr.astype(np.float32)
 
 
+def apply_windowing(image: np.ndarray, subset: str) -> np.ndarray:
+    """Apply CT windowing based on the dataset.
+
+    For CT images, applies window level and width transformation.
+    For MRI images (windowing=None), returns the image unchanged.
+
+    Args:
+        image: Raw image array (e.g., in Hounsfield Units for CT)
+        subset: Dataset subset name to determine windowing parameters
+
+    Returns:
+        Windowed image array
+    """
+    windowing = DEFAULT_WINDOWINGS.get(subset)
+
+    # No windowing for MRI or unknown datasets
+    if windowing is None:
+        return image
+
+    window_level = windowing["window_level"]
+    window_width = windowing["window_width"]
+
+    # Apply CT windowing transformation
+    # Convert window level/width to min/max values
+    window_min = window_level - window_width / 2
+    window_max = window_level + window_width / 2
+
+    # Clip and normalize to [0, 1] range
+    windowed = np.clip(image, window_min, window_max)
+    windowed = (windowed - window_min) / (window_max - window_min)
+
+    return windowed.astype(np.float32)
+
+
 def to_display_image(image: np.ndarray) -> np.ndarray:
     """Normalise image for display purposes (uint8, 3-channel)."""
 
@@ -416,7 +450,11 @@ def load_dataset_sample(
     try:
         target_id = parse_target_selection(target_selection)
         image, caption, meta = sample_dataset_example(subset, target_id)
-        display, mask_msg = render_image_with_mask_info(image, meta.get("mask"))
+
+        # Apply windowing only for display, keep raw image for model inference
+        windowed_image = apply_windowing(image, subset)
+        display, mask_msg = render_image_with_mask_info(windowed_image, meta.get("mask"))
+
         target = meta.get("target")
         meta_text = caption
         if target is not None:
@@ -439,7 +477,7 @@ def load_dataset_sample(
             meta_text,
             pd.DataFrame(),
             ground_truth_update,
-            {"image": image, "mask": meta.get("mask")},
+            {"image": image, "mask": meta.get("mask")},  # Store raw image for inference
         )
     except Exception as exc:  # pragma: no cover - surfaced in UI
         return None, f"Failed to load sample: {exc}", pd.DataFrame(), gr.update(visible=False), None
@@ -463,19 +501,24 @@ def run_inference(
 
 def handle_upload_preview(
     image: np.ndarray | Image.Image | None,
+    subset: str,
 ) -> Tuple[Optional[np.ndarray], str, pd.DataFrame, Dict[str, Any], Optional[Dict[str, Any]]]:
     if image is None:
         return None, "Please upload an image.", pd.DataFrame(), gr.update(visible=False), None
 
     try:
         np_image = to_numpy_image(image)
-        display = to_display_image(np_image)
+
+        # Apply windowing only for display, keep raw image for model inference
+        windowed_image = apply_windowing(np_image, subset)
+        display = to_display_image(windowed_image)
+
         return (
             display,
             "Image uploaded. Click 'Run inference' to compute predictions.",
             pd.DataFrame(),
             gr.update(visible=False),
-            {"image": np_image, "mask": None},
+            {"image": np_image, "mask": None},  # Store raw image for inference
         )
     except Exception as exc:  # pragma: no cover - surfaced in UI
         return None, f"Failed to load image: {exc}", pd.DataFrame(), gr.update(visible=False), None