Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

ADSPAO: Enhanced artemisinin optimization for multi-threshold segmentation of chronic obstructive pulmonary disease.

iScience·2026
Same author

Evolutionary-based deep learning network model using adaptive mixing differential evolution and application in acute pulmonary embolism.

Journal of advanced research·2026
Same author

An optimized machine learning model based on hematological indicators for the noninvasive identification of baicalin's therapeutic effects in pulmonary hypertension.

Computer methods in biomechanics and biomedical engineering·2026
Same author

Prey capture enhanced Harris hawks optimizer for wrapper-based feature selection in high-dimensional medical data.

Computer methods and programs in biomedicine·2026
Same author

Gaussian bare‑bone JAYA algorithm for multi-threshold medical image segmentation.

Scientific reports·2025
Same author

Enhanced kernel search algorithm for optimizing local search capability and its application to carbon fiber draft process.

PloS one·2025

Related Experiment Video

Updated: Oct 17, 2025

A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment
10:39

A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment

Published on: May 24, 2022

2.5K

Performance optimization of differential evolution with slime mould algorithm for multilevel breast cancer image

Lei Liu1, Dong Zhao1, Fanhua Yu1

  • 1College of Computer Science and Technology, Changchun Normal University, Changchun, Jilin, 130032, China.

Computers in Biology and Medicine
|October 12, 2021
PubMed
Summary

This study introduces a new computational method to improve how breast cancer images are segmented. By combining two nature-inspired algorithms, the researchers created a tool that accurately identifies different tissue types in medical scans, helping doctors diagnose the disease more effectively.

Keywords:
Breast cancerDifferential evolutionImage segmentationOptimizationSlime mould algorithmSwarm-intelligenceslime mould algorithmmultilevel image segmentationevolutionary computationpathological image analysis

Frequently Asked Questions

More Related Videos

Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates
11:24

Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates

Published on: March 7, 2017

7.1K
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
07:15

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

7.0K

Related Experiment Videos

Last Updated: Oct 17, 2025

A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment
10:39

A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment

Published on: May 24, 2022

2.5K
Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates
11:24

Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates

Published on: March 7, 2017

7.1K
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
07:15

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

7.0K

Area of Science:

  • Medical imaging informatics within diagnostic radiology
  • Computational intelligence and modified differential evolution optimization

Background:

Breast cancer remains a significant threat to global female health, necessitating advanced diagnostic support tools. Medical image processing serves as a vital component of modern clinical diagnostic workflows. Image segmentation represents the primary stage for extracting meaningful information from complex medical scans. Multilevel segmentation techniques offer a direct and efficient approach for analyzing these intricate visual data sets. Recent efforts have applied evolutionary and population-based strategies to automate this segmentation process. However, these existing approaches frequently suffer from limited convergence precision and often become trapped in local optima. No prior work had successfully resolved these persistent computational limitations in breast cancer image analysis. This gap motivated the development of a more robust optimization framework for medical image segmentation.

Purpose Of The Study:

The study aims to develop a robust optimization method for multilevel breast cancer image segmentation. Researchers sought to address the persistent issues of poor convergence accuracy in existing evolutionary algorithms. The project specifically targets the tendency of current population-based methods to become trapped in local optima. By drawing inspiration from slime mould foraging behaviors, the authors intended to create a more efficient search strategy. The team also aimed to build a high-quality segmentation model based on non-local means 2D histograms. This effort was motivated by the need for reliable diagnostic assistance in medical imaging. The authors focused on validating their new approach through both mathematical benchmarks and clinical image sets. Ultimately, the work strives to provide a practical tool for the analysis of invasive ductal carcinoma pathological images.

Main Methods:

The review approach involved developing a hybrid optimization framework inspired by biological foraging patterns. Researchers integrated slime mould behaviors into the standard differential evolution algorithm to enhance search capabilities. The team constructed a multilevel segmentation model using non-local means 2D histograms and 2D Kapur's entropy. They utilized IEEE CEC 2014 benchmark functions to evaluate the mathematical robustness of the new algorithm. Initial validation occurred through applying the model to a standardized reference image set. The study then compared the proposed method against several peer evolutionary algorithms. Finally, the authors tested the model on clinical breast invasive ductal carcinoma images to assess real-world applicability. This structured testing sequence ensured both algorithmic stability and practical diagnostic utility.

Main Results:

Key findings from the literature indicate that the modified differential evolution algorithm achieves higher convergence accuracy than traditional population-based methods. The experimental results demonstrate that the proposed model effectively avoids local optima during the segmentation process. Quantitative comparisons on IEEE CEC 2014 benchmarks confirm the superior optimization performance of the modified approach. Testing on reference images shows that the model produces high-quality segmentation outputs suitable for medical analysis. When applied to breast invasive ductal carcinoma images, the method provides clear and accurate tissue differentiation. The data suggest that the hybrid model consistently outperforms peer algorithms in both speed and precision. These results validate the efficacy of incorporating slime mould foraging logic into evolutionary computation. The findings confirm that the developed model serves as a robust tool for pathological image processing tasks.

Conclusions:

The authors demonstrate that their modified approach achieves superior convergence accuracy compared to existing evolutionary methods. This synthesis suggests that integrating slime mould foraging behaviors effectively mitigates the risk of becoming trapped in local optima. The findings imply that the proposed model provides a reliable framework for segmenting complex medical imagery. The researchers conclude that their technique offers practical utility for analyzing breast invasive ductal carcinoma pathological images. This study highlights the potential of hybridizing population-based algorithms to enhance diagnostic image processing capabilities. The evidence indicates that the model maintains high-quality segmentation performance across diverse test sets. These results support the broader application of the modified differential evolution model in clinical imaging research. The authors emphasize that their work provides a foundation for future improvements in automated diagnostic assistance systems.

The researchers propose a hybrid model that integrates slime mould foraging behaviors into a differential evolution framework. This mechanism enhances convergence precision and prevents the algorithm from stalling in local optima, unlike standard population-based methods which often struggle with these specific computational hurdles during image processing tasks.

The model utilizes a non-local means 2D histogram alongside 2D Kapur's entropy. These components are necessary to define the objective function, allowing the algorithm to categorize pixel intensities effectively while maintaining spatial consistency in breast cancer images compared to simpler intensity-based thresholding techniques.

The authors utilize IEEE CEC 2014 benchmark functions to verify the algorithm's optimization capabilities. This technical necessity ensures the modified approach performs reliably before testing on actual medical data, contrasting with studies that apply new algorithms directly to clinical images without rigorous mathematical validation.

The 2D histogram acts as the primary data structure for spatial information. It plays a role in capturing local pixel relationships, which is more robust against noise than traditional 1D histograms used in earlier segmentation models, providing a clearer distinction between cancerous and healthy tissue regions.

The researchers measure convergence accuracy and the ability to avoid local optima. These metrics quantify the algorithm's success, showing that the modified differential evolution approach consistently outperforms standard peers when processing invasive ductal carcinoma images, which are notoriously difficult to segment due to high textural complexity.

The authors propose that their model provides practical support for future pathological image analysis. They suggest that this high-quality segmentation method could assist clinicians in diagnostic tasks, contrasting with previous, less accurate models that failed to provide the necessary precision for reliable clinical decision-making.