Learning Gaussian Mixture-distributed Prototypes for 3D Scene Graph Generation from RGB-D Sequences

Rongxing Ding, Hongyu Qu, Xinguang Xiang, Pengpeng Li, Xiangbo Shu
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:25114-25129, 2026.

Abstract

3D Scene Graph Generation (3DSGG) aims to create a structured representation of 3D environment by identifying objects as nodes and their relations as edges. Existing 3DSGG methods based on RGB-D sequences typically put much focus on the adaption of neural networks to robust node and edge feature extraction in complex 3D scenes, yet ignoring the inherent intra-class diversity within each class and inter-class similarity between different categories associated with nodes and edges. In this work, we develop GMPSSG, a novel Gaussian Mixture-distributed Prototype mining framework for 3DSGG. Specifically, we model different categories with independent Gaussian Mixture-distributed Prototype to effectively mitigate inter-class similarity, while employing multiple Gaussian components within each prototype to capture intra-class diversity. Moreover, Prototype-anchored Representation Learning is introduced to construct a well-structured and mutually independent category space; Topology-aware Prototype Interaction is devised to capture implicit co-occurrence priors within the scene, and leverage them to calibrate prototype distributions, thereby ensuring the plausibility of node-edge matching. Experiments on 3DSSG dataset demonstrate GMPSSG outperforms various top-leading methods. Our code is available at GMPSSG.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-ding26n, title = {Learning {G}aussian Mixture-distributed Prototypes for 3{D} Scene Graph Generation from {RGB}-D Sequences}, author = {Ding, Rongxing and Qu, Hongyu and Xiang, Xinguang and Li, Pengpeng and Shu, Xiangbo}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {25114--25129}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/ding26n/ding26n.pdf}, url = {https://proceedings.mlr.press/v306/ding26n.html}, abstract = {3D Scene Graph Generation (3DSGG) aims to create a structured representation of 3D environment by identifying objects as nodes and their relations as edges. Existing 3DSGG methods based on RGB-D sequences typically put much focus on the adaption of neural networks to robust node and edge feature extraction in complex 3D scenes, yet ignoring the inherent intra-class diversity within each class and inter-class similarity between different categories associated with nodes and edges. In this work, we develop GMPSSG, a novel Gaussian Mixture-distributed Prototype mining framework for 3DSGG. Specifically, we model different categories with independent Gaussian Mixture-distributed Prototype to effectively mitigate inter-class similarity, while employing multiple Gaussian components within each prototype to capture intra-class diversity. Moreover, Prototype-anchored Representation Learning is introduced to construct a well-structured and mutually independent category space; Topology-aware Prototype Interaction is devised to capture implicit co-occurrence priors within the scene, and leverage them to calibrate prototype distributions, thereby ensuring the plausibility of node-edge matching. Experiments on 3DSSG dataset demonstrate GMPSSG outperforms various top-leading methods. Our code is available at GMPSSG.} }
Endnote
%0 Conference Paper %T Learning Gaussian Mixture-distributed Prototypes for 3D Scene Graph Generation from RGB-D Sequences %A Rongxing Ding %A Hongyu Qu %A Xinguang Xiang %A Pengpeng Li %A Xiangbo Shu %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-ding26n %I PMLR %P 25114--25129 %U https://proceedings.mlr.press/v306/ding26n.html %V 306 %X 3D Scene Graph Generation (3DSGG) aims to create a structured representation of 3D environment by identifying objects as nodes and their relations as edges. Existing 3DSGG methods based on RGB-D sequences typically put much focus on the adaption of neural networks to robust node and edge feature extraction in complex 3D scenes, yet ignoring the inherent intra-class diversity within each class and inter-class similarity between different categories associated with nodes and edges. In this work, we develop GMPSSG, a novel Gaussian Mixture-distributed Prototype mining framework for 3DSGG. Specifically, we model different categories with independent Gaussian Mixture-distributed Prototype to effectively mitigate inter-class similarity, while employing multiple Gaussian components within each prototype to capture intra-class diversity. Moreover, Prototype-anchored Representation Learning is introduced to construct a well-structured and mutually independent category space; Topology-aware Prototype Interaction is devised to capture implicit co-occurrence priors within the scene, and leverage them to calibrate prototype distributions, thereby ensuring the plausibility of node-edge matching. Experiments on 3DSSG dataset demonstrate GMPSSG outperforms various top-leading methods. Our code is available at GMPSSG.
APA
Ding, R., Qu, H., Xiang, X., Li, P. & Shu, X.. (2026). Learning Gaussian Mixture-distributed Prototypes for 3D Scene Graph Generation from RGB-D Sequences. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:25114-25129 Available from https://proceedings.mlr.press/v306/ding26n.html.

Related Material