Kezheng Xiong

Kezheng Xiong

he/him

3D computer vision researcher, a Ph.D. student in Xiamen University

Publications

You can also find my articles on my Google Scholar profile.

Conference Papers

MoT3DVG: A Benchmark for Outdoor 3D Visual Grounding with Motion-Aware Descriptions and Temporal Cues

Published in NeurIPS 2026 ED Track, 2026

We therefore introduce MoT3DVG, a large-scale dataset for dynamic-aware outdoor 3DVG with temporally evolving motion descriptions. It contains 850 scenarios with 31,128 frames from nuScenes dataset, and provides 144,568 language prompts with motion-aware descriptions for dynamic objects across time. We further propose DynaVG, a novel framework that effectively leverages temporal cues for outdoor 3DVG.

Citation

Coming Soon

Download Paper

Unsupervised Point Cloud Registration via Training-Time Semantic Guidance

Published in ECCV 2026, 2026

We propose CAESAR, a teacher-student framework guided by an off-the-shelf 3D segmentation model exclusively during training.

Citation

Xiong, K., Xu, S., Ao, S., Shen, S., Wang, C., Wen, C. (2026). Unsupervised Point Cloud Registration via Training-Time Semantic Guidance. In: Favaro, P., Kukelova, Z., Maki, A., Rohrbach, A., Schindler, K., Tombari, F. (eds) Computer Vision – ECCV 2026. ECCV 2026. Lecture Notes in Computer Science, vol 17035. Springer, Cham. https://doi.org/10.1007/978-3-032-37464-6_26

Download Paper

TACO: Task-Aware Contrastive Learning for Joint LiDAR Localization and 3D Object Detection

Published in CVPR 2026, 2026

This paper proposes TACO, the first Task-Aware COntrastive learning framework, which performs joint LiDAR localization and 3D object detection within a single, unified network.

Citation

Leyuan, Xing, et al. "TACO: Task-Aware Contrastive Learning for Joint LiDAR Localization and 3D Object Detection", In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Download Paper

Mining and Transferring Feature-Geometry Coherence for Unsupervised Point Cloud Registration

Published in NeurIPS 2024 Main Conference, 2024

In this paper, we propose a novel unsupervised registration method termed INTEGER to incorporate high-level contextual information for reliable pseudo-label mining.

Citation

Xiong, Kezheng, et al. "Mining and Transferring Feature-Geometry Coherence for Unsupervised Point Cloud Registration." The Thirty-eighth Annual Conference on Neural Information Processing Systems (2024).

Download Paper

SPEAL: Skeletal Prior Embedded Attention Learning for Cross-Source Point Cloud Registration

Published in AAAI-24 Main Track, 2024

In this paper, we propose a novel method termed SPEAL to leverage skeletal representations for effective learning of intrinsic topologies of point clouds, facilitating robust capture of geometric intricacy.

Citation

Xiong, Kezheng, et al. "SPEAL: Skeletal Prior Embedded Attention Learning for Cross-Source Point Cloud Registration." Proceedings of the AAAI Conference on Artificial Intelligence. Vol. 38. No. 6. 2024.

Download Paper