Related Experiment Video
Updated: Aug 24, 2025
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
6.9K
Immunity-based Ebola optimization search algorithm for minimization of feature extraction with reduction in digital
Olaide N Oyelade1, Absalom E Ezugwu2
1School of Mathematics, Statistics, and Computer Science, University of KwaZulu-Natal, King Edward Avenue, Pietermaritzburg Campus, Pietermaritzburg, 3201, KwaZulu-Natal, South Africa.
Scientific Reports
|October 26, 2022
Summary
This study introduces an optimized feature selection method for convolutional neural networks (CNNs) to improve breast cancer detection in mammography. The novel immunity-based Ebola optimization search algorithm (IEOSA) enhances classification accuracy by selecting discriminant features.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computational Optimization
Background:
- Feature classification in digital medical images, particularly mammography, is an optimization challenge often overlooked.
- Convolutional Neural Networks (CNNs) excel at feature extraction and classification for disease detection, but suboptimal feature selection can hinder performance.
Purpose of the Study:
- To address the performance limitations in CNN-based medical image classification caused by unoptimized feature selection.
- To develop a metaheuristic algorithm for optimizing the number of features extracted by CNNs, ensuring only relevant features are used for classification.
Main Methods:
- A novel variant of the Ebola-based optimization algorithm, incorporating population immunity and chaos mapping initialization, termed the immunity-based Ebola optimization search algorithm (IEOSA).
- Application of IEOSA to optimize feature selection by processing noisy features from CNN convolutional layers.
- Evaluation of IEOSA using benchmark functions and comparison with existing optimization algorithms.
Main Results:
- IEOSA demonstrated strong performance on classical and IEEE CEC benchmarked functions.
- The algorithm effectively enhanced feature selection for CNNs in digital mammography.
- The IEOSA method significantly improved classification accuracy in breast cancer prediction using CNN models.
Conclusions:
- The proposed IEOSA is an effective metaheuristic algorithm for feature optimization in CNNs.
- IEOSA successfully addresses the challenge of selecting discriminant features, leading to improved classification performance.
- This approach offers a promising advancement for accurate breast cancer detection in digital mammography.
