Scholarlink

<p>The interval type-2 fuzzy clustering (IT2FCM) algorithm is an effective method for data clustering and image segmentation. However, the traditional IT2FCM-based algorithms cannot achieve accurate segmentation results on noisy images. To improve the accuracy and robustness of the IT2FCM-based algorithms for noisy images, this paper proposes a novel adaptive interval type-2 fuzzy clustering algorithm with local and non-local spatial information. The main contributions of the proposed algorithm are as follows: (1) a new weighted local fuzzy factor embedding non-local spatial information of the image is devised; (2) two regularizing factors are introduced into the new weighted local fuzzy factor; (3) the distances in the objective function are corrected by using non-local spatial information; and (4) the new weighted local fuzzy factor is fused with the interval type-2 fuzzy C-means clustering algorithm. Using these innovative tricks and an effective type reduction method, the proposed algorithm can achieve higher performance and more accurate segmentation results of noisy images than existing several algorithms.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Adaptive Interval Type-2 Fuzzy Clustering Noisy Image Segmentation Algorithm with Weighted Local Spatial Information Embedding Non-local Spatial Information

  • 期刊论文

  • 发表日期：

    2025年7月19日

查看全文
  • Chengquan Huang,
  • Huan Lei,
  • Jialei Peng,
  • Xiaosu Qin,
  • Yang Chen,
  • Lihua Zhou

摘要

The interval type-2 fuzzy clustering (IT2FCM) algorithm is an effective method for data clustering and image segmentation. However, the traditional IT2FCM-based algorithms cannot achieve accurate segmentation results on noisy images. To improve the accuracy and robustness of the IT2FCM-based algorithms for noisy images, this paper proposes a novel adaptive interval type-2 fuzzy clustering algorithm with local and non-local spatial information. The main contributions of the proposed algorithm are as follows: (1) a new weighted local fuzzy factor embedding non-local spatial information of the image is devised; (2) two regularizing factors are introduced into the new weighted local fuzzy factor; (3) the distances in the objective function are corrected by using non-local spatial information; and (4) the new weighted local fuzzy factor is fused with the interval type-2 fuzzy C-means clustering algorithm. Using these innovative tricks and an effective type reduction method, the proposed algorithm can achieve higher performance and more accurate segmentation results of noisy images than existing several algorithms.