Category-Specific Progressive Refinement for Multimodal 3D Object Detection
Applied Sciences · 19.11.2026 · 16 (22) · 11467
DOI: 10.3390/app162211467 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 19.11.2026. Verfasst von Shiwei Huang und 5 weiteren Autor:innen. Die Arbeit nennt 25 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Open Project Program of the National Key Laboratory of Uranium Resources Exploration—Mining and Nuclear Remote Sensing, Autonomous Deployment Project of the National Key Laboratory of Uranium Resources Exploration—Mining and Nuclear Remote Sensing, Natural Science Foundation of Jiangxi Province.
Abstract (Originalsprache)
Residual errors in multimodal 3D object detection can arise at different stages and vary across object categories. We propose a category-specific progressive refinement framework for a CenterHead–LoGoHead detector using RGB, thermal, and LiDAR inputs. Pedestrian Proposal Rescue (PPR) supplements first-stage pedestrian candidates through constrained proposal merging. Hard-Mismatch Geometry Refinement (HMR) predicts seven-dimensional residual corrections for pedestrian boxes using supervision segmented by proposal IoU. Scale-Routed Pyramid RoI Refinement (SPR) enhances Cyclist and Vehicle region-of-interest (RoI) features by combining multiple aggregation resolutions within a fixed physical region, with scale weights conditioned on category, distance, and response occupancy. Experiments use the 5637-frame DSERT-RoLL validation split. The original progressive model achieves 69.90% mAP and 62.17% mAPH, exceeding the seed-666 Base detector by 6.25 and 4.03 percentage points, respectively. Across seeds 666, 123, and 2026, the IoU-extended progressive configuration achieves 69.23 ± 0.09% mAP and 61.76 ± 0.06% mAPH. Joint HMR and SPR optimization from the seed-2026 PPR checkpoint achieves 69.93% mAP and 62.35% mAPH. Standalone module ablations and analyses across weather and lighting conditions characterize the contributions and remaining weaknesses of the refinement pipeline. The results support category- and stage-specific refinement within an existing tri-modal detection architecture.
Förderung
- Open Project Program of the National Key Laboratory of Uranium Resources Exploration—Mining and Nuclear Remote Sensing
- Autonomous Deployment Project of the National Key Laboratory of Uranium Resources Exploration—Mining and Nuclear Remote Sensing
- Natural Science Foundation of Jiangxi Province
- National Natural Science Foundation of China
Zitieren
APA
Huang, S., Xie, X., Xu, G., Yu, M., Yang, H., & Xia, Y. (2026). Category-Specific Progressive Refinement for Multimodal 3D Object Detection. Applied Sciences, 16(22), 11467. https://doi.org/10.3390/app162211467
Vancouver
Huang S, Xie X, Xu G, Yu M, Yang H, Xia Y. Category-Specific Progressive Refinement for Multimodal 3D Object Detection. Applied Sciences. 2026;16(22):11467. doi:10.3390/app162211467
BibTeX
@article{Huang2026,
author = {Huang, Shiwei and Xie, Xiaowei and Xu, Guangyu and Yu, Mei and Yang, Hongzhi and Xia, Yuanping},
title = {Category-Specific Progressive Refinement for Multimodal 3D Object Detection},
journal = {Applied Sciences},
year = {2026},
volume = {16},
number = {22},
pages = {11467},
doi = {10.3390/app162211467}
}
EUVIMED