EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Hierarchical Compositional Alignment for Zero-Shot Part-Level Segmentation

Shan Yang iD, Shujie Ji, Zhendong Xiao, Xiongding Liu iD, Wu Wei iD

Sensors · 30.03.2026 · 26 (7) · 2130

DOI: 10.3390/s26072130 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 30.03.2026. Verfasst von Shan Yang und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Automation Science and Engineering, South China University of Technology, Guangzhou 510640, China. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Science and Technology Planning Project of Guangdong Province, China, Zhejiang Provincial Natural Science Foundation of China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In robotic fine-grained tasks (e.g., grasping and assembly), precise interaction requires a detailed understanding of object components. While Visual Language Models (VLMs) excel at object-level recognition, they struggle with part-level segmentation (e.g., knife handles), limiting performance in complex scenarios. VLMs face three key challenges: (1) Visual granularity mismatch—object-level features lack part-level details; (2) Semantic hierarchy gaps—parts and objects differ significantly in semantics; (3) Cross-modal bias—CLIP’s text–image alignment favors global over local features. To address these, we propose a one-stage VLM-based part segmentation method. First, the Hierarchy-Aware Feature Selection mechanism analyzes Transformer features in different hierarchies to enhance spatial and semantic precision for part segmentation. Second, the Multi-Hierarchy Feature Adapter bridges object-to-part feature granularity via the hierarchical adaptation. Finally, the Hierarchical Multimodal Alignment Module harmonizes classification accuracy and mask integrity via hierarchical alignment of vision–language, mitigating the bias of CLIP’s object-level priori knowledge. Experiments show the proposed method improves part segmentation performance for Zero-Shot, achieving 25.86% on Pascal-Part and 13.09% on ADE20K-Part (gains of +0.81% hIoU and +2.96% hIoU over baseline). This work advances robotic visual perception, with applications in intelligent manufacturing and intelligent service.

Schlagwörter

Stichwörter der Autor:innen

multi-hierarchy featuremultimodal alignmentpart segmentationvisual language models

Autor:innen und Institutionen

  • School of Automation Science and Engineering, South China University of Technology, Guangzhou 510640, China.
  • School of Automation, Hangzhou Dianzi University, Hangzhou 310018, China.

ORCID-Profile: Shan Yang, Xiongding Liu, Wu Wei

Förderung

  • National Natural Science Foundation of China
  • Science and Technology Planning Project of Guangdong Province, China
  • Zhejiang Provincial Natural Science Foundation of China
  • Fundamental Research Funds for the Provincial Universities of Zhejiang under Grant
  • Fundamental Research Project of Hangzhou Dianzi University
  • National Natural Science Foundation of China https://ror.org/01h0zpd94
  • Zhejiang Provincial Natural Science Foundation of China https://ror.org/04aqat463
  • Fundamental Research Funds for the Provincial Universities of Zhejiang

Zitieren

APA

Yang, S., Ji, S., Xiao, Z., Liu, X., & Wei, W. (2026). Hierarchical Compositional Alignment for Zero-Shot Part-Level Segmentation. Sensors, 26(7), 2130. https://doi.org/10.3390/s26072130

Vancouver

Yang S, Ji S, Xiao Z, Liu X, Wei W. Hierarchical Compositional Alignment for Zero-Shot Part-Level Segmentation. Sensors. 2026;26(7):2130. doi:10.3390/s26072130

BibTeX
@article{Yang2026,
  author = {Yang, Shan and Ji, Shujie and Xiao, Zhendong and Liu, Xiongding and Wei, Wu},
  title = {Hierarchical Compositional Alignment for Zero-Shot Part-Level Segmentation},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {7},
  pages = {2130},
  doi = {10.3390/s26072130}
}
Kostenlos starten →