README.md

Modified SceneFun3D Dataset for Open Vocabulary Functional 3D Scene Graphs

In evaluation of OpenFunGraph, we do not use the newest released version of SceneFun3D. Here we released the version we used and the annotations we added on this version.

Annotations

SceneFun3D.annotations.json: Object and interactive element segmentation annotations

SceneFun3D.relations.json: Functional 3D scene graph annotations

all_labels.json: all labels appeared in the dataset for evaluation

all_labels_clip.embedding.npy: CLIP embeddings of all labels appeared in the dataset for evaluation

all_edges.json: all relationship descriptions appeared in the dataset for edge evaluation

all_edges_bert_embeddings.npy: BERT embeddings of all relationship descriptions appeared in the dataset for edge evaluation

OpenFunGraph_split.txt: split used for OpenFunGraph evaluation

Assets for each scene

Here the assets and the usage are the same with SceneFun3D (https://scenefun3d.github.io/documentation/). We only use part of the dataset's assets for our work.

highres_depth: the ground-truth depth image projected from the mesh generated by Faro’s laser scanners (1920x1440 in landscape mode, 1440x1920 in portrait mode) - 10 FPS

wide: high resolution RGB images of the wide camera (1920x1440 in landscape mode, 1440x1920 in portrait mode) - 10 FPS

wide_intrinsics: camera intrinsics for the high resolution images

lowres_wide.traj: contains the ARKit camera pose trajectory in the ARKit coordinate system

metadata.csv: information about landscape or portrait for each scene

xxx_refined_transform.npy: 4x4 transformation matrix that registers the Faro laser scan to the ARKit coordinate system

xxx_laser_scan.ply: combined Faro laser scan with 5mm resolution