Precision and efficiency in skin cancer segmentation through a dual encoder deep learning model
Abstract Skin cancer is a prevalent health concern, and accurate segmentation of skin lesions is crucial for early diagnosis. Existing methods for skin lesion segmentation often face trade-offs between efficiency and feature extraction capabilities. This paper proposes Dual Skin Segmentation (DuaSki...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-02-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-88753-3 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823862505684860928 |
|---|---|
| author | Asaad Ahmed Guangmin Sun Anas Bilal Yu Li Shouki A. Ebad |
| author_facet | Asaad Ahmed Guangmin Sun Anas Bilal Yu Li Shouki A. Ebad |
| author_sort | Asaad Ahmed |
| collection | DOAJ |
| description | Abstract Skin cancer is a prevalent health concern, and accurate segmentation of skin lesions is crucial for early diagnosis. Existing methods for skin lesion segmentation often face trade-offs between efficiency and feature extraction capabilities. This paper proposes Dual Skin Segmentation (DuaSkinSeg), a deep-learning model, to address this gap by utilizing dual encoders for improved performance. DuaSkinSeg leverages a pre-trained MobileNetV2 for efficient local feature extraction. Subsequently, a Vision Transformer-Convolutional Neural Network (ViT-CNN) encoder-decoder architecture extracts higher-level features focusing on long-range dependencies. This approach aims to combine the efficiency of MobileNetV2 with the feature extraction capabilities of the ViT encoder for improved segmentation performance. To evaluate DuaSkinSeg’s effectiveness, we conducted experiments on three publicly available benchmark datasets: ISIC 2016, ISIC 2017, and ISIC 2018. The results demonstrate that DuaSkinSeg achieves competitive performance compared to existing methods, highlighting the potential of the dual encoder architecture for accurate skin lesion segmentation. |
| format | Article |
| id | doaj-art-aa963473af38411e8a6fb8ff91ad573f |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-aa963473af38411e8a6fb8ff91ad573f2025-02-09T12:30:21ZengNature PortfolioScientific Reports2045-23222025-02-0115111610.1038/s41598-025-88753-3Precision and efficiency in skin cancer segmentation through a dual encoder deep learning modelAsaad Ahmed0Guangmin Sun1Anas Bilal2Yu Li3Shouki A. Ebad4School of Information Science and Technology, Beijing University of TechnologySchool of Information Science and Technology, Beijing University of TechnologyCollege of Information Science and Technology, Hainan Normal UniversitySchool of Information Science and Technology, Beijing University of TechnologyCenter for Scientific Research and Entrepreneurship, Northern Border UniversityAbstract Skin cancer is a prevalent health concern, and accurate segmentation of skin lesions is crucial for early diagnosis. Existing methods for skin lesion segmentation often face trade-offs between efficiency and feature extraction capabilities. This paper proposes Dual Skin Segmentation (DuaSkinSeg), a deep-learning model, to address this gap by utilizing dual encoders for improved performance. DuaSkinSeg leverages a pre-trained MobileNetV2 for efficient local feature extraction. Subsequently, a Vision Transformer-Convolutional Neural Network (ViT-CNN) encoder-decoder architecture extracts higher-level features focusing on long-range dependencies. This approach aims to combine the efficiency of MobileNetV2 with the feature extraction capabilities of the ViT encoder for improved segmentation performance. To evaluate DuaSkinSeg’s effectiveness, we conducted experiments on three publicly available benchmark datasets: ISIC 2016, ISIC 2017, and ISIC 2018. The results demonstrate that DuaSkinSeg achieves competitive performance compared to existing methods, highlighting the potential of the dual encoder architecture for accurate skin lesion segmentation.https://doi.org/10.1038/s41598-025-88753-3Skin Lesion SegmentationDual EncoderVision Transformer (ViT)Convolutional Neural Networks (CNNs)ViT-CNN |
| spellingShingle | Asaad Ahmed Guangmin Sun Anas Bilal Yu Li Shouki A. Ebad Precision and efficiency in skin cancer segmentation through a dual encoder deep learning model Scientific Reports Skin Lesion Segmentation Dual Encoder Vision Transformer (ViT) Convolutional Neural Networks (CNNs) ViT-CNN |
| title | Precision and efficiency in skin cancer segmentation through a dual encoder deep learning model |
| title_full | Precision and efficiency in skin cancer segmentation through a dual encoder deep learning model |
| title_fullStr | Precision and efficiency in skin cancer segmentation through a dual encoder deep learning model |
| title_full_unstemmed | Precision and efficiency in skin cancer segmentation through a dual encoder deep learning model |
| title_short | Precision and efficiency in skin cancer segmentation through a dual encoder deep learning model |
| title_sort | precision and efficiency in skin cancer segmentation through a dual encoder deep learning model |
| topic | Skin Lesion Segmentation Dual Encoder Vision Transformer (ViT) Convolutional Neural Networks (CNNs) ViT-CNN |
| url | https://doi.org/10.1038/s41598-025-88753-3 |
| work_keys_str_mv | AT asaadahmed precisionandefficiencyinskincancersegmentationthroughadualencoderdeeplearningmodel AT guangminsun precisionandefficiencyinskincancersegmentationthroughadualencoderdeeplearningmodel AT anasbilal precisionandefficiencyinskincancersegmentationthroughadualencoderdeeplearningmodel AT yuli precisionandefficiencyinskincancersegmentationthroughadualencoderdeeplearningmodel AT shoukiaebad precisionandefficiencyinskincancersegmentationthroughadualencoderdeeplearningmodel |