Refactor Cutout and Image Editing Logic

- Renamed "Base Plane Depth" to "Reference Depth" across validate_cutout.py for consistency.
- Introduced a new specification for Cutout mesh sampling to ensure source pixel spacing is preserved and limits are applied correctly.
- Unified naming conventions for image editing results to maintain consistency across Remove Background, Upscale, and Crop operations.
- Updated the handling of Refine Selection to ensure it retains the original bounds and does not apply unnecessary cropping.
- Enhanced material generation logic to connect image color directly to IOR Level, improving the rendering of dark areas.
- Established clear requirements for geometry depth controls and image result materials to ensure consistent behavior across different components.
- Removed outdated logic and ensured all changes are documented and tested for regression.

src/anyimage/server/jobs/cutout.py

@@ -1,6 +1,5 @@
 """Generate optional AI files from one rectangular Cutout image input."""
 
-import json
 from pathlib import Path
 
 from ..geometry.prediction_artifacts import generate_moge2_artifacts
@@ -15,22 +14,14 @@ def run(context, parameters):
     result = {}
     if parameters.get("refine_selection"):
         context.progress(0.2, "Refining Selection")
-        selection_image_path, bounds = refine_selection(
+        selection_image_path = refine_selection(
             context,
             input_path,
             parameters["refine_model"],
             parameters["device"],
             context.directory / "selection.png",
         )
-        bounds_path = context.directory / "selection_bounds.json"
-        bounds_path.write_text(
-            json.dumps(bounds, separators=(",", ":")),
-            encoding="utf-8",
-        )
-        result.update(
-            selection_image=selection_image_path.name,
-            selection_bounds=bounds_path.name,
-        )
+        result["selection_image"] = selection_image_path.name
 
     depth_mode = str(parameters.get("depth_mode", "NONE"))
     normal_mode = str(parameters.get("normal_mode", "NONE"))

src/anyimage/server/jobs/refine_image_selection.py

@@ -1,27 +1,24 @@
-import json
 import time
 
 from ..selection import refine_selection
 
 
 def run(context, parameters):
-    """Refine a gesture mask with BEN2 and return its source-image bounds."""
+    """Refine foreground Alpha while preserving the submitted bounds canvas."""
     started = time.perf_counter()
     context.progress(0.05, "Preparing image Selection")
     context.check_cancelled()
     output_directory = context.directory
     output_directory.mkdir(parents=True, exist_ok=True)
     context.progress(0.2, f"Loading BEN2 for {parameters['device'].upper()}")
-    output_path, bounds = refine_selection(
+    output_path = refine_selection(
         context,
         parameters["input"],
         parameters["model"],
         parameters["device"],
         output_directory / "image.png",
     )
-    bounds_path = output_directory / "bounds.json"
-    bounds_path.write_text(json.dumps(bounds, separators=(",", ":")), encoding="utf-8")
     elapsed_ms = (time.perf_counter() - started) * 1000.0
     message = f"Image Selection refined in {elapsed_ms:.0f} ms"
     context.progress(1.0, message)
-    return {"selection_image": output_path.name, "selection_bounds": bounds_path.name}
+    return {"selection_image": output_path.name}

src/anyimage/server/selection/refine_selection.py

@@ -1,24 +1,6 @@
 from ..models.ben2 import infer_foreground
 
 
-def _crop_to_alpha(image):
-    import numpy as np
-
-    pixels = np.asarray(image.convert("RGBA"), dtype=np.uint8)
-    visible = pixels[:, :, 3] > 0
-    visible_y = np.flatnonzero(np.any(visible, axis=1))
-    visible_x = np.flatnonzero(np.any(visible, axis=0))
-    if not len(visible_x):
-        raise ValueError("The generated Selection contains no visible pixels")
-    bounds = (
-        int(visible_x[0]),
-        int(visible_y[0]),
-        int(visible_x[-1]) + 1,
-        int(visible_y[-1]) + 1,
-    )
-    return image.crop(bounds), bounds
-
-
 def refine_selection(
     context,
     input_path,
@@ -37,6 +19,5 @@ def refine_selection(
         preserve_input_alpha=True,
     )
     context.check_cancelled()
-    output, bounds = _crop_to_alpha(foreground)
-    output.save(output_path, format="PNG", compress_level=1)
-    return output_path, bounds
+    foreground.convert("RGBA").save(output_path, format="PNG", compress_level=1)
+    return output_path