article · IEEE Access
This research proposes a deep learning model for the accurate classification of skin lesions, addressing the high mortality rate associated with melanoma and the diagnostic challenges posed by similar-looking lesions. The model employs transfer learning with a pre-trained GoogleNet architecture, fine-tuning its parameters through training. It was tested using the ISIC 2019 public challenge dataset and successfully classified eight distinct types of skin lesions, including melanoma, basal cell carcinoma, and squamous cell carcinoma. The model achieved a classification accuracy of 94.92%, a sensitivity of 79.8%, a specificity of 97%, and a precision of 80.36%. Additionally, it can identify images that do not belong to any of the eight specified classes, classifying them as unknown.
Accurate and early diagnosis of skin lesions, particularly melanoma, is vital for effective treatment and saving lives. This research offers a highly accurate automated tool that could assist dermatologists in distinguishing between various skin conditions, potentially improving diagnostic efficiency and patient outcomes in a critical area of healthcare.
This research presents an applied deep learning model for automated skin lesion classification, which could serve as a diagnostic aid for dermatologists and medical professionals. The model's ability to classify eight lesion types and identify unknown images suggests its potential for integration into clinical decision support systems. It is an early-stage technology, tested on a public dataset, indicating its readiness for further validation and development towards real-world clinical application.
AI-generated from the published abstract. Always read the original work before citing.
Melanoma is a type of skin cancer with a high mortality rate. The different types of skin lesions result in an inaccurate diagnosis due to their high similarity. Accurate classification of the skin lesions in their early stages enables dermatologists to treat the patients and save their lives. This paper proposes a model for a highly accurate classification of skin lesions. The proposed model utilized the transfer learning and pre-trained model with GoogleNet. The model parameters are used as initial values, and then these parameters will be modified through training. The latest well-known public challenge dataset, ISIC 2019, is used to test the ability of the proposed model to classify different kinds of skin lesions. The proposed model successfully classified the eight different classes of skin lesions, namely, melanoma, melanocytic nevus, basal cell carcinoma, actinic keratosis, benign keratosis, dermatofibroma, vascular lesion, and Squamous cell carcinoma. The achieved classification accuracy, sensitivity, specificity, and precision percentages are 94.92%, 79.8%, 97%, and 80.36%, respectively. The proposed model can detect images that do not belong to any one of the eight classes where these images are classified as unknown images.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1109/access.2020.3003890
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
