Datasets:

FrameCut3D / README.md
CCRss's picture
Card: real and gso configs, GSO section
16e6297 verified
|
Raw History Blame Contribute Delete
3.37 kB
---
license: cc-by-nc-4.0
pretty_name: FrameCut3D
task_categories:
- image-to-3d
tags:
- 3d-generation
- gaussian-splatting
- benchmark
- truncation
size_categories:
- n<1K
configs:
- config_name: real
data_dir: real
default: true
- config_name: gso
data_dir: synthetic/gso
---
# FrameCut3D
FrameCut3D is a benchmark for **out-of-frame truncation** in single-image 3D generators: the object in the input
photo runs off the image frame, and the generator silently returns a truncated shape. It accompanies the paper
"FrameCut3D: Diagnosing Out-of-Frame Completion in Single-Image 3D Generators" (under review).
## Real split (66 objects, `real/`)
Everyday objects captured with a handheld phone orbit. Each object folder contains:
| File | Description |
|---|---|
| `object.png` | Isolated full-object image (background removed), the complete input. |
| `cuts/cut40.png` ... `cut70.png` | Controlled truncation: the bottom 40 to 70% of the object's bounding box is masked. |
| `frame.jpg` | The full capture frame the object image was taken from. |
| `crop.png`, `crop.json` | Genuine frame truncation: the fixed center crop of the capture, with its severity and cut sides (6 objects are not cut). |
| `measured_geometry.ply` | Reference geometry from structure-from-motion and multi-view stereo on the orbit, support plane removed (coloured points). |
| `pseudo_gt.ply` | The generator's reconstruction from the complete image (TripoSplat Gaussians), used as pseudo ground truth. |
| `video.mp4` | The capture orbit (1080p, 30 fps). |
| `meta.json` | Label, description, category, geometry tier (1 solid to 4 reflective or transparent), symmetry flag, common/uncommon category flag. |
The measured geometry misses part of the surface next to the support plane. The paper quantifies this bias.
Coordinates are in the reconstruction frame (arbitrary scale and rotation); evaluation normalizes each shape to a
unit box and searches the rotation.
## Synthetic splits (`synthetic/`)
- **GSO held-out split (`synthetic/gso/`, viewer config `gso`)**: 200 Google Scanned Objects, each rendered from three
views (azimuth 0, 120, 240 degrees, elevation 20 degrees, 1024x1024 RGBA). Per object and view: `input.png` and
`cuts/cut40.png` ... `cut70.png` (bottom 40 to 70% of the object's bounding box made transparent). Per object:
`surface.ply` with 50,000 ground-truth surface points in the render frame. GSO is CC BY 4.0 (Google LLC).
- `gso_200.json`: the same object list with categories and view settings.
- `objaverse_48.json`: the 48 Objaverse objects of the main synthetic split (UIDs, one view at azimuth 180,
elevation 20). Objaverse assets keep their original licences and are not redistributed here.
## Scripts (`scripts/`)
`truncate.py` (cut rule and hidden band) and `bandmetrics.py` (Chamfer, F-score, hidden-band recall after alignment).
## Load
```python
from datasets import load_dataset
real = load_dataset("issai/FrameCut3D", "real", split="train") # 66 real objects
gso = load_dataset("issai/FrameCut3D", "gso", split="train") # 200 GSO objects x 3 views
```
## Licence
CC BY-NC 4.0 for the real split and the annotations. GSO renders and surface points follow the GSO licence (CC BY 4.0, Google LLC); Objaverse assets keep their source licences.
## Citation
The paper is under review. A citation will be added on publication.