Implement BiRefNet and HAT Sharper models with background removal capabilities

- Added validation documentation for AI model tiers.
- Introduced `background.py` for background removal logic, including inference functions for alpha and foreground extraction.
- Created `onnx_birefnet.py` to handle BiRefNet model sessions and alpha inference.
- Developed tests for background models, ensuring proper dispatch and session reuse.
- Added tests for BiRefNet ONNX model to validate alpha output and error handling for invalid values.
- Implemented model export scripts for BiRefNet and HAT Sharper, ensuring proper weight compression and licensing.
- Created `deform_sampling.py` for deformable convolution operations used in BiRefNet.
- Added utility functions for compressing ONNX weights without altering graph arithmetic.
- Introduced a reference plane node for depth field adjustments in geometry processing.

tests/anyimage/operators/test_ai_image_operators.py

@@ -22,7 +22,7 @@ class AiImageOperatorsTest(BlenderTestCase):
         operator = operator_type()
         operator.input_path = "input.png"
         operator._image_target = SimpleNamespace(start_frame=4)
-        operator.model = "SAFMN_REAL_X4"
+        operator.model = "HAT_GAN_X4_SHARPER"
         operator.max_input_size = 1024
         context = SimpleNamespace(
             scene=SimpleNamespace(anyimage_settings=SimpleNamespace(video_frames=7))
@@ -37,12 +37,12 @@ class AiImageOperatorsTest(BlenderTestCase):
         ):
             request = operator.request(context)
 
-        require_model.assert_called_once_with("SAFMN_REAL_X4")
+        require_model.assert_called_once_with("HAT_GAN_X4_SHARPER")
         self.assertEqual(
             request,
             {
                 "input": "input.png",
-                "model": "SAFMN_REAL_X4",
+                "model": "HAT_GAN_X4_SHARPER",
                 "device": "cuda",
                 "video_frames": 7,
                 "start_frame": 4,
@@ -108,12 +108,12 @@ class AiImageOperatorsTest(BlenderTestCase):
         result = SimpleNamespace(directory=Path("result"))
         operator = self.upscale.UpscaleImage()
         operator._image_target = Mock()
-        operator.model = "SAFMN_REAL_X4"
+        operator.model = "HAT_GAN_X4_SHARPER"
         with (
             patch.object(
                 self.upscale,
                 "upscale_model_label",
-                return_value="SAFMN Real x4",
+                return_value="HAT Sharper",
             ),
         ):
             message = operator.response(None, result)
@@ -122,4 +122,4 @@ class AiImageOperatorsTest(BlenderTestCase):
             Path("result"),
             frame_prefix="upscale.",
         )
-        self.assertEqual(message, "Image is upscaled with SAFMN Real x4")
+        self.assertEqual(message, "Image is upscaled with HAT Sharper")