Optimized Training for Convolutional Neural Network Using Enhanced Grey Wolf Optimization Algorithm

Convolutional Neural Networks (CNNs) are widely used in image classification tasks and have achieved significant performance. They have different applications with great success, especially in the medical field. The choice of architecture and hyperparameter settings of the CNN, highly influences bot...

Full description

Saved in:
Bibliographic Details
Published in:Informatica (Ljubljana) 2021-07, Vol.45 (5), p.731-739
Main Authors: Guernine, Akram, Kimour, Mohamed Tahar
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Artificial intelligence
Artificial neural networks
Convergence
Deep learning
Design optimization
Genetic algorithms
Image classification
Image processing
Medical imaging
Medical research
Neural networks
Optimization algorithms
Skin cancer
Spatial resolution
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Convolutional Neural Networks (CNNs) are widely used in image classification tasks and have achieved significant performance. They have different applications with great success, especially in the medical field. The choice of architecture and hyperparameter settings of the CNN, highly influences both accuracy and its convergence speed. The empirical design and optimization of a new CNN architecture require a lot of expertise and can be very time-consuming. This paper proposes an enhanced Grey Wolf Optimization (GWO) algorithm to efficiently explore a defined space of potentially suitable CNN architectures, and simultaneously optimize their hyperparameters. Moreover, we introduce a spatial resolution reduction for a given image processing task, while taking skin cancer detection as a practical application. Through conducted experiments, we have shown that the obtained results are better than other classification methods in terms of accuracy and convergence speed.
ISSN:0350-5596
1854-3871
DOI:10.31449/inf.v45i5.3497