feat: GPU-side ImageNet normalize + concat for full-frame path

PIL bicubic vs MLX cubic diff too large (0.22-0.59) — resize stays
on CPU. Only normalize+concat moved to MLX. Tiled path unchanged.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

src/corridorkey_mlx/engine.py

@@ -26,6 +26,7 @@ from corridorkey_mlx.io.image import (
     postprocess_foreground,
     preprocess,
 )
+from corridorkey_mlx.io.preprocess_mlx import preprocess_mlx
 
 if TYPE_CHECKING:
     from corridorkey_mlx.model.corridorkey import GreenFormer
@@ -178,7 +179,7 @@ class CorridorKeyMLXEngine:
                 )
                 mask_f32 = np.asarray(mask_pil, dtype=np.float32)[:, :, np.newaxis] / 255.0
 
-            x = preprocess(rgb_f32, mask_f32)
+            x = preprocess_mlx(rgb_f32, mask_f32)
             outputs = self._model(x)
             # NOTE: mx.eval is MLX array materialization, not Python eval()
             mx.eval(outputs)  # noqa: S307

src/corridorkey_mlx/io/preprocess_mlx.py

@@ -0,0 +1,41 @@
+"""GPU-side preprocessing for full-frame inference.
+
+Moves ImageNet normalization and channel concatenation to MLX while
+keeping PIL bicubic resize on CPU (MLX cubic kernel differs too much).
+"""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+import mlx.core as mx
+
+if TYPE_CHECKING:
+    import numpy as np
+
+# ImageNet normalization constants (GPU-resident)
+_IMAGENET_MEAN = mx.array([0.485, 0.456, 0.406])
+_IMAGENET_STD = mx.array([0.229, 0.224, 0.225])
+
+
+def preprocess_mlx(rgb_f32: np.ndarray, mask_f32: np.ndarray) -> mx.array:
+    """GPU-side normalize + concat for full-frame inference.
+
+    Assumes resize already happened on CPU via PIL bicubic.
+
+    Args:
+        rgb_f32: (H, W, 3) float32 in [0, 1], already resized.
+        mask_f32: (H, W, 1) float32 in [0, 1], already resized.
+
+    Returns:
+        (1, H, W, 4) mx.array — ImageNet-normalized RGB + raw alpha hint.
+    """
+    img = mx.array(rgb_f32)
+    mask = mx.array(mask_f32)
+
+    # ImageNet normalize on GPU
+    img = (img - _IMAGENET_MEAN) / _IMAGENET_STD
+
+    # Concat and add batch dim
+    combined = mx.concatenate([img, mask], axis=-1)  # (H, W, 4)
+    return mx.expand_dims(combined, axis=0)  # (1, H, W, 4)