Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Oct 24;19(10):e0309430.
doi: 10.1371/journal.pone.0309430. eCollection 2024.

A Multi-level ensemble approach for skin lesion classification using Customized Transfer Learning with Triple Attention

Anwar Hossain Efat  1 , S M Mahedy Hasan  1 , Md Palash Uddin  2 , Md Al Mamun  1
Affiliations

A Multi-level ensemble approach for skin lesion classification using Customized Transfer Learning with Triple Attention

Anwar Hossain Efat et al. PLoS One. .

Abstract

Skin lesions encompass a variety of skin abnormalities, including skin diseases that affect structure and function, and skin cancer, which can be fatal and arise from abnormal cell growth. Early detection of lesions and automated prediction is crucial, yet accurately identifying responsible regions post-dominance dispersion remains a challenge in current studies. Thus, we propose a Convolutional Neural Network (CNN)-based approach employing a Customized Transfer Learning (CTL) model and Triple Attention (TA) modules in conjunction with Ensemble Learning (EL). While Ensemble Learning has become an integral component of both Machine Learning (ML) and Deep Learning (DL) methodologies, a specific technique ensuring optimal allocation of weights for each model's prediction is currently lacking. Consequently, the primary objective of this study is to introduce a novel method for determining optimal weights to aggregate the contributions of models for achieving desired outcomes. We term this approach "Information Gain Proportioned Averaging (IGPA)," further refining it to "Multi-Level Information Gain Proportioned Averaging (ML-IGPA)," which specifically involves the utilization of IGPA at multiple levels. Empirical evaluation of the HAM1000 dataset demonstrates that our approach achieves 94.93% accuracy with ML-IGPA, surpassing state-of-the-art methods. Given previous studies' failure to elucidate the exact focus of black-box models on specific regions, we utilize the Gradient Class Activation Map (GradCAM) to identify responsible regions and enhance explainability. Our study enhances both accuracy and interpretability, facilitating early diagnosis and preventing the consequences of neglecting skin lesion detection, thereby addressing issues related to time, accessibility, and costs.

PubMed Disclaimer

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1. Sample Images of Different Classes (a) Actinic Keratosis (b) Basal Cell Carcinoma (c) Benign Keratosis (d) Dermatofibroma (e) Melanoma (f) Nevus (g) Vascular Lesions.
Fig 2
Fig 2. Instance distribution for each class.
Fig 3
Fig 3. Schematic representation of methodology.
Fig 4
Fig 4. Sample images of the images after augmentation.
(a) Original Sample, (b) Rotated Sample, (c) Width_shifted, (d) Height_shifted, (e) Zoomed, (f), Horizontal_flipped, (g) Vertical_flipped.
Fig 5
Fig 5. CTL architecture.
Fig 6
Fig 6. Feature extraction after activation of each layer (One image as example).
Fig 7
Fig 7. IGPA in Level—n.
Fig 8
Fig 8. Organization of ML-IGPA.
Fig 9
Fig 9. Confusion matrix obtained by SL-IGPA.
Fig 10
Fig 10. Confusion matrix obtained by ML-IGPA (All classifiers).
Fig 11
Fig 11. Confusion matrix obtained by ML-IGPA (Best 3 classifiers).
Fig 12
Fig 12. ROC-AUC curve obtained by SL-IGPA.
Fig 13
Fig 13. ROC-AUC curve obtained by ML-IGPA (All classifiers).
Fig 14
Fig 14. ROC-AUC curve obtained by ML-IGPA (Best 3 classifiers).
Fig 15
Fig 15. GradCAM generation by the model for each class.
(a) GradCAM for AK, (b) GradCAM for BCC, (c) GradCAM for BKL, (d) GradCAM for DF, (e) GradCAM for MEL, (f) GradCAM for NV, (g) GradCAM for VASC.
Fig 16
Fig 16. GradCAM visualization for architecture explainability (Example by DN121).
(a) Original, (b) CACNN, (c) SEACNN, (d) SACNN.
See this image and copyright information in PMC

References

    1. Bibi S, Khan MA, Shah JH, Damaševičius R, Alasiry A, Marzougui M, et al. MSRNet: multiclass skin lesion recognition using additional residual block based fine-tuned deep models information fusion and best feature selection. Diagnostics. 2023. Sep 26;13(19):3063. doi: 10.3390/diagnostics13193063 - DOI - PMC - PubMed
    1. Dillshad V, Khan MA, Nazir M, Saidani O, Alturki N, Kadry S. D2LFS2Net: Multi‐class skin lesion diagnosis using deep learning and variance‐controlled Marine Predator optimisation: An application for precision medicine. CAAI Transactions on Intelligence Technology. 2023. doi: 10.1049/cit2.12267 - DOI
    1. Hussain M, Khan MA, Damaševičius R, Alasiry A, Marzougui M, Alhaisoni M, et al. SkinNet-INIO: multiclass skin lesion localization and classification using fusion-assisted deep neural networks and improved nature-inspired optimization algorithm. Diagnostics. 2023. Sep 6;13(18):2869. doi: 10.3390/diagnostics13182869 - DOI - PMC - PubMed
    1. Ahmad N, Shah JH, Khan MA, Baili J, Ansari GJ, Tariq U, et al. A novel framework of multiclass skin lesion recognition from dermoscopic images using deep learning and explainable AI. Frontiers in Oncology. 2023. Jun 6;13:1151257. doi: 10.3389/fonc.2023.1151257 - DOI - PMC - PubMed
    1. Malik S, Akram T, Awais M, Khan MA, Hadjouni M, Elmannai H, et al. An improved skin lesion boundary estimation for enhanced-intensity images using hybrid metaheuristics. Diagnostics. 2023. Mar 28;13(7):1285. doi: 10.3390/diagnostics13071285 - DOI - PMC - PubMed