Add PXDepth demo

app.py

@@ -193,41 +193,21 @@ def filter_flying_points(
     return (points[keep], colors[keep]) if keep.any() else (points, colors)
 
 
-def write_splat(
+def write_viewer_glb(
     path: Path,
     points: np.ndarray,
     colors: np.ndarray,
-    point_size: float,
 ) -> None:
-    """Write a compact Gaussian-splat viewer file with adjustable point size."""
+    """Write the browser point cloud using the stable GLB viewer path."""
+    import trimesh
+
     display_points = points * np.array([1.0, -1.0, -1.0], dtype=np.float32)
-    low, high = np.percentile(display_points, [1.0, 99.0], axis=0)
-    extent = max(float(np.max(high - low)), 1e-6)
-    spacing = extent / max(float(np.sqrt(points.shape[0])), 1.0)
-    scale = max(spacing * float(point_size) * 0.65, 1e-7)
-
-    dtype = np.dtype(
-        [
-            ("position", "<f4", (3,)),
-            ("scale", "<f4", (3,)),
-            ("color", "u1", (4,)),
-            ("rotation", "u1", (4,)),
-        ],
-        align=False,
-    )
-    splats = np.empty(points.shape[0], dtype=dtype)
-    splats["position"] = display_points
-    splats["scale"] = scale
-    splats["color"][:, :3] = colors
-    splats["color"][:, 3] = 255
-    splats["rotation"] = np.array([255, 128, 128, 128], dtype=np.uint8)
-    splats.tofile(path)
+    trimesh.PointCloud(display_points, colors=colors).export(path)
 
 
 def update_viewer(
     cache_path: Optional[str],
     filter_points: bool,
-    point_size: float,
     max_points: int,
 ) -> Optional[str]:
     """Rebuild the viewer from cached points without running either model."""
@@ -244,10 +224,10 @@ def update_viewer(
         raise gr.Error("No points remain after filtering.")
 
     cache_file = Path(cache_path)
-    tag = f"{int(max_points)}_{int(filter_points)}_{float(point_size):.2f}"
-    viewer_path = cache_file.with_name(f"pointcloud_viewer_{tag}.splat")
-    write_splat(viewer_path, points, colors, point_size)
-    for old_path in cache_file.parent.glob("pointcloud_viewer_*.splat"):
+    tag = f"{int(max_points)}_{int(filter_points)}"
+    viewer_path = cache_file.with_name(f"pointcloud_viewer_{tag}.glb")
+    write_viewer_glb(viewer_path, points, colors)
+    for old_path in cache_file.parent.glob("pointcloud_viewer_*.*"):
         if old_path != viewer_path:
             old_path.unlink(missing_ok=True)
     return str(viewer_path)
@@ -293,7 +273,6 @@ def on_submit(
     image: Optional[np.ndarray],
     apply_mask: bool,
     filter_points: bool,
-    point_size: float,
     max_points: int,
     request: gr.Request,
 ):
@@ -351,7 +330,6 @@ def on_submit(
     viewer_path = update_viewer(
         str(cache_path),
         filter_points,
-        point_size,
         max_points,
     )
 
@@ -380,6 +358,7 @@ See the [paper](https://arxiv.org/abs/2608.16984),
                         label="Input Image",
                         image_mode="RGB",
                         type="numpy",
+                        placeholder="# Drop an image here\n— or —\nClick to upload",
                         elem_id="img-display-input",
                     )
                     with gr.Accordion(label="Settings", open=False):
@@ -392,14 +371,6 @@ See the [paper](https://arxiv.org/abs/2608.16984),
                             info="Statistical outlier filtering; does not rerun the model.",
                             value=False,
                         )
-                        point_size = gr.Slider(
-                            0.25,
-                            3.0,
-                            value=1.0,
-                            step=0.05,
-                            label="Point Size",
-                            info="Updates only the 3D viewer.",
-                        )
                         max_points = gr.Slider(
                             50_000,
                             500_000,
@@ -449,24 +420,18 @@ See the [paper](https://arxiv.org/abs/2608.16984),
 
             submit.click(
                 on_submit,
-                [input_image, apply_mask, filter_points, point_size, max_points],
+                [input_image, apply_mask, filter_points, max_points],
                 [depth_map, model_3d, downloads, viewer_cache],
                 show_progress="full",
                 concurrency_limit=1,
             )
-            viewer_inputs = [viewer_cache, filter_points, point_size, max_points]
+            viewer_inputs = [viewer_cache, filter_points, max_points]
             filter_points.change(
                 update_viewer,
                 viewer_inputs,
                 model_3d,
                 show_progress="minimal",
             )
-            point_size.release(
-                update_viewer,
-                viewer_inputs,
-                model_3d,
-                show_progress="minimal",
-            )
             max_points.release(
                 update_viewer,
                 viewer_inputs,

requirements.txt

@@ -8,5 +8,6 @@ huggingface_hub>=0.28
 matplotlib>=3.8
 pillow>=10.0
 scipy>=1.12
+trimesh>=4.0
 git+https://github.com/EasternJournalist/utils3d.git@3fab839f0be9931dac7c8488eb0e1600c236e183
 git+https://github.com/microsoft/MoGe.git@42acd8f46e974f5e0548ecd72315d1d7df2cb6f4