Construction Algorithm for Adaptive Morphological Structural Elements Based on Fuzzy C-Means Clustering with a Greyscale Fluctuation Factor
Year: 2026
Type: article
Abstract: Conventional morphological operations commonly utilize a structural element with a single shape and fixed size for image processing. Due to the complexity and diversity of image targets, morphological operations using this structural element may lead to changes in the image target feature information. For example, the target boundary contours may be shifted, new artificial targets may appear, and target details may be lost. To address this, a morphological adaptive structural element construction method based on fuzzy C-means (FCM) is proposed. The constructed structural elements are able to change their shapes adaptively according to the edge features of the image target. First, a new FCM objective function is constructed utilizing the grey fluctuation factor to improve the homogeneity of pixel clustering. Then, using the objective function and incorporating the density peaking (DP) algorithm, a local subregion of the image is clustered. Finally, the boundary points between the classes after clustering are considered to be the adaptive structural elements of this subregion. The experimental results show that morphological operations on an image that adopt this adaptive structural element can obtain the adaptive structural element corresponding to the shape of the image target, maintain the integrity of the necessary information of the target, avoid a target boundary contour offset, and improve the accuracy of target boundary contour localization. In addition, this method effectively preserves the geometric structural properties of the image and reduces the distortion level of the target structural details of the image.
Source: Mathematics
DOI: 10.3390/math14224223
Authors: Chao Fang, Wang Xiaopeng, Tongyi Wei
Institution: Lanzhou Jiaotong University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,137
Funder: Lanzhou Jiaotong University
Award: G9001395213
OpenAlex ID: W4297881110
