Papers › Spatiality-guided Transformer for 3D Dense Captioning on Point Clouds

Spatiality-guided Transformer for 3D Dense Captioning on Point Clouds

22 Apr 2022arXiv:2204.10688archive 2025-07-28

Heng Wang, Chaoyi Zhang, Jianhui Yu, Weidong Cai

Dense captioning in 3D point clouds is an emerging vision-and-language task involving object-level 3D scene understanding. Apart from coarse semantic class prediction and bounding box regression as in traditional 3D object detection, 3D dense captioning aims at producing a further and finer instance-level label of natural language description on visual appearance and spatial relations for each scene object of interest. To detect and describe objects in a scene, following the spirit of neural machine translation, we propose a transformer-based encoder-decoder architecture, namely SpaCap3D, to transform objects into descriptions, where we especially investigate the relative spatiality of objects in 3D scenes and design a spatiality-guided encoder via a token-to-token spatial relation learning objective and an object-centric decoder for precise and spatiality-enhanced object caption generation. Evaluated on two benchmark datasets, ScanRefer and ReferIt3D, our proposed SpaCap3D outperforms the baseline method Scan2Cap by 4.94% and 9.61% in CIDEr@0.5IoU, respectively. Our project page with source code and supplementary files is available at https://SpaCap3D.github.io/ .

Paper PDF Code

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

heng-hw/spacap3d officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

3D Object Detection 3D dense captioning Caption Generation Decoder Dense Captioning Machine Translation Object Object Detection Scene Understanding object-detection

Results from the paper archive 2025-07-28

Task	Dataset	Model	Metric	Value	Rank at snapshot	Leaderboard	Report
3D dense captioning	Nr3D	SpaCap3d	BLEU-4	19.92	#9 of 10	Archive leaderboard	report
3D dense captioning	Nr3D	SpaCap3d	CIDEr	33.71	#9 of 10	Archive leaderboard	report
3D dense captioning	Nr3D	SpaCap3d	METEOR	22.61	#9 of 10	Archive leaderboard	report
3D dense captioning	Nr3D	SpaCap3d	ROUGE-L	50.50	#9 of 10	Archive leaderboard	report
3D dense captioning	ScanRefer Dataset	SpaCap3d	BLEU-4	35.30	#8 of 12	Archive leaderboard	report
3D dense captioning	ScanRefer Dataset	SpaCap3d	CIDEr	58.06	#8 of 12	Archive leaderboard	report
3D dense captioning	ScanRefer Dataset	SpaCap3d	METEOR	26.16	#8 of 12	Archive leaderboard	report
3D dense captioning	ScanRefer Dataset	SpaCap3d	ROUGE-L	55.03	#8 of 12	Archive leaderboard	report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position Encodings Adam Attention BPE Dense Connections Dropout Label Smoothing Layer Normalization Linear Layer Multi-Head Attention Position-Wise Feed-Forward Layer Residual Connection Softmax Transformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections