DeepLabv3+-Based Segmentation and Best Features Selection Using Slime Mould Algorithm for Multi-Class Skin Lesion Classification
Mathematics · 2023-01-10 · 11 (2) · 364
DOI: 10.3390/math11020364 ↗At a glance
Journal article in Mathematics, published online on 2023-01-10. Written by Mehwish Zafar and 5 co-authors. Cited 53 times so far (Crossref); 96 references. The full text is freely available (open access, CC BY 4.0).
Abstract (original language)
The development of abnormal cell growth is caused by different pathological alterations and some genetic disorders. This alteration in skin cells is very dangerous and life-threatening, and its timely identification is very essential for better treatment and safe cure. Therefore, in the present article, an approach is proposed for skin lesions’ segmentation and classification. So, in the proposed segmentation framework, pre-trained Mobilenetv2 is utilised in the act of the back pillar of the DeepLabv3+ model and trained on the optimum parameters that provide significant improvement for infected skin lesions’ segmentation. The multi-classification of the skin lesions is carried out through feature extraction from pre-trained DesneNet201 with N × 1000 dimension, out of which informative features are picked from the Slim Mould Algorithm (SMA) and input to SVM and KNN classifiers. The proposed method provided a mean ROC of 0.95 ± 0.03 on MED-Node, 0.97 ± 0.04 on PH2, 0.98 ± 0.02 on HAM-10000, and 0.97 ± 0.00 on ISIC-2019 datasets.
Cite
APA
Zafar, M., Amin, J., Sharif, M., Anjum, M. A., Mallah, G. A., & Kadry, S. (2023). DeepLabv3+-Based Segmentation and Best Features Selection Using Slime Mould Algorithm for Multi-Class Skin Lesion Classification. Mathematics, 11(2), 364. https://doi.org/10.3390/math11020364
Vancouver
Zafar M, Amin J, Sharif M, Anjum MA, Mallah GA, Kadry S. DeepLabv3+-Based Segmentation and Best Features Selection Using Slime Mould Algorithm for Multi-Class Skin Lesion Classification. Mathematics. 2023;11(2):364. doi:10.3390/math11020364
BibTeX
@article{Zafar2023,
author = {Zafar, Mehwish and Amin, Javeria and Sharif, Muhammad and Anjum, Muhammad Almas and Mallah, Ghulam Ali and Kadry, Seifedine},
title = {DeepLabv3+-Based Segmentation and Best Features Selection Using Slime Mould Algorithm for Multi-Class Skin Lesion Classification},
journal = {Mathematics},
year = {2023},
volume = {11},
number = {2},
pages = {364},
doi = {10.3390/math11020364}
}
EUVIMED