Upload ActionBench dataset (imgs + surfaces + cameras)

README.md

@@ -26,17 +26,21 @@ viewer: false
 ## 📖 Overview
 
 ActionBench is a benchmark dataset of **128 paired video ↔ animated point-cloud samples** for evaluating animated 3D mesh generation from video.
+The dataset consists of synthetic scenes of animated objects from [ObjaverseXL](https://objaverse.allenai.org/), rendered using **Blender 3.5.1**.
 
 Each sample contains:
 - **Video**: 16 RGBA frames with alpha mask
+- **Camera** (`camera.json`): Camera parameters using Blender convention (`X_cam = X @ R^T + T`, camera looks along -Z). See [`projection.py`](projection.py) for how to project the point cloud onto the image plane.
 - **Animated Point Cloud**: Surface points sampled on the animated object with shape `(T, V, 6)` where:
   - `T=16`: number of keyframes
-  - `V`: number of vertices (points randomly sampled on the mesh surface)
+  - `V=100_000`: number of vertices (points randomly sampled on the mesh surface)
   - `6`: position `(x, y, z)` + normal `(nx, ny, nz)` for each point
 
   > **Note:** The point cloud is **tracked**: each point index corresponds to the same surface point deformed across timesteps, providing dense correspondences over time.
 
-The dataset consists of synthetic scenes of animated objects from [ObjaverseXL](https://objaverse.allenai.org/), rendered using **Blender 3.5.1**.
+  The animation lie in normalized space `[-1., 1.]^3`.
+
+
 
 ## 📊 Evaluation
 

data/000-000_03b69da8d2c94b5999bcf2605ee2ecd9/camera.json

@@ -0,0 +1,32 @@
+{
+  "R": [
+    [
+      0.7071067690849304,
+      0.7071067690849304,
+      3.057652975257952e-08
+    ],
+    [
+      -0.06162842735648155,
+      0.06162835657596588,
+      0.9961947202682495
+    ],
+    [
+      0.704416036605835,
+      -0.704416036605835,
+      0.08715569972991943
+    ]
+  ],
+  "T": [
+    -7.968946806848997e-09,
+    -8.911828860565645e-08,
+    -2.990313554974874
+  ],
+  "focal_length_ndc": [
+    2.1875,
+    2.1875
+  ],
+  "principal_point_ndc": [
+    0.0,
+    0.0
+  ]
+}

data/000-000_03b69da8d2c94b5999bcf2605ee2ecd9/surfaces.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:993065dc051f9148b1804cf4f08e09cad469721b38318730c8527c9d61c1abbc
+oid sha256:8584ac060305f54e71d7937f2611b10de9e7a3b296625d8752c6f787ee2cc284
 size 38400128

data/000-000_f15298421b3d4e0fab4c43863a7e72fd/camera.json

@@ -0,0 +1,32 @@
+{
+  "R": [
+    [
+      0.7071067690849304,
+      0.7071067690849304,
+      3.057652975257952e-08
+    ],
+    [
+      -0.06162842735648155,
+      0.06162835657596588,
+      0.9961947202682495
+    ],
+    [
+      0.704416036605835,
+      -0.704416036605835,
+      0.08715569972991943
+    ]
+  ],
+  "T": [
+    -7.990723424103153e-09,
+    -8.936182076973652e-08,
+    -2.9984851384143116
+  ],
+  "focal_length_ndc": [
+    2.1875,
+    2.1875
+  ],
+  "principal_point_ndc": [
+    0.0,
+    0.0
+  ]
+}

data/000-000_f15298421b3d4e0fab4c43863a7e72fd/surfaces.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:72def77950803ad1824dade1dfe02704eab8f1c754447a61d1d33ba2dd8cae52
+oid sha256:3432d2e0760461ee3dc4d4fae5b18f4bd1fd03ac3b7f38757f3c3444dfcc2ddc
 size 38400128

data/000-001_095641b042f049c3aa86792836f8993d/camera.json

@@ -0,0 +1,32 @@
+{
+  "R": [
+    [
+      0.7071067690849304,
+      0.7071067690849304,
+      3.057652975257952e-08
+    ],
+    [
+      -0.06162842735648155,
+      0.06162835657596588,
+      0.9961947202682495
+    ],
+    [
+      0.704416036605835,
+      -0.704416036605835,
+      0.08715569972991943
+    ]
+  ],
+  "T": [
+    -7.963818829512475e-09,
+    -8.906094143352491e-08,
+    -2.988389303187194
+  ],
+  "focal_length_ndc": [
+    2.1875,
+    2.1875
+  ],
+  "principal_point_ndc": [
+    0.0,
+    0.0
+  ]
+}