+Advanced Search
Research on Segmentation Algorithm for GH4169 Alloy Grains Based on a Joint Attention Mechanism
Author:
Affiliation:

Western Superconducting Technologies Co., Ltd, Xi'an 710018, China

Clc Number:

TG132.3+3

Fund Project:

陕西省重点研发计划(项目号2025CY-YBXM-237)

  • Article
  • |
  • Figures
  • |
  • Metrics
  • |
  • Reference
  • |
  • Related
  • |
  • Cited by
  • |
  • Materials
  • |
  • Comments
    Abstract:

    To tackle the segmentation challenges arising from the complex morphology, significant size variations, blurred boundaries, and tight interconnections of superalloy grains, this research presents a grain segmentation network integrated with a joint suppression attention mechanism that fuses channel and spatial information. The proposed network combines the global modeling capacity of Swin Transformer and the local detail restoration capability of a convolutional neural network (CNN), and embeds the aforementioned joint suppression attention mechanism, which integrates channel and spatial information, into the decoder. Results show that this mechanism effectively suppresses noise and texture interference, enhances the abilities of feature screening and generalization, and reinforces the fusion of shallow and deep features, thereby markedly improving the continuity of grain boundaries. Experimental results demonstrate that the proposed algorithm achieves an IoU of 67.34% and an F1-score of 78.62% on the self-constructed metallographic dataset, with all metrics outperforming those of mainstream grain segmentation algorithms for superalloys.

    Reference
    Related
    Cited by
Get Citation

[Deng Bowen, Li Ke, Zhang Xinyuan, Zhao Yanru, Shi Puying, Wang Hu, Gao Huixian, Yang Chao. Research on Segmentation Algorithm for GH4169 Alloy Grains Based on a Joint Attention Mechanism[J]. Rare Metal Materials and Engineering,2026,55(10):2565~2572.]
DOI:10.12442/j. issn.1002-185X.20250360

Copy
Article Metrics
  • Abstract:
  • PDF:
  • HTML:
  • Cited by:
History
  • Received:July 08,2025
  • Revised:September 05,2025
  • Adopted:September 08,2025
  • Online: August 24,2026
  • Published: July 31,2026