Related Experiment Video
Updated: Oct 19, 2025
12:27
Large-scale Reconstructions and Independent, Unbiased Clustering Based on Morphological Metrics to Classify Neurons in Selective Populations
Published on: February 15, 2017
7.1K
Morphological Reconstruction-Based Image-Guided Fuzzy Clustering with a Novel Impact Factor
Qingxue Qin1, Guangmei Xu1,2, Jin Zhou1
1Shandong Provincial Key Laboratory of Network Based Intelligent Computing, University of Jinan, Jinan 250022, China.
Journal of Healthcare Engineering
|September 24, 2021
Summary
This study introduces improved fuzzy c-means (FCM) clustering methods using a guided filter for better noisy image segmentation. The new approaches enhance adaptability and accuracy, especially for heavily corrupted images.
Area of Science:
- Computer Vision
- Image Processing
- Machine Learning
Background:
- Guided filter enhances image smoothing and edge preservation.
- Existing guided filter-FCM methods struggle with adaptability due to fixed parameters.
- Noisy images present challenges for accurate image segmentation.
Purpose of the Study:
- To develop adaptive guided filter-FCM methods for improved noisy image segmentation.
- To enhance the robustness and accuracy of FCM clustering on images with varying noise levels.
- To simplify parameter selection for guided filter-based FCM algorithms.
Main Methods:
- Proposed an improved FCM with Guided Filter (IFCM_GF) using an adjustable influence factor (ρ) for the guidance image.
- Introduced a Morphological Reconstruction-based Improved FCM with Guided Filter (MRIFCM_GF) for heavily noisy images.
- Implemented morphological reconstruction (MR) to pre-process images before IFCM_GF application.
Main Results:
- IFCM_GF demonstrated excellent segmentation results on various noisy images by dynamically adjusting the influence factor.
- MRIFCM_GF achieved superior segmentation accuracy on heavily noisy images compared to IFCM_GF.
- MRIFCM_GF offers a simplified selection of the influence factor and effective noise removal.
Conclusions:
- The proposed IFCM_GF and MRIFCM_GF methods significantly improve fuzzy c-means clustering for noisy image segmentation.
- MRIFCM_GF is particularly effective for images with heavy noise due to the integration of morphological reconstruction.
- These adaptive methods offer enhanced performance and user-friendliness for image analysis tasks.
