EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Progenesis of Skin Disease Detection Using Deep Learning

Firoz Akhtar, Dr. K. P. Yadav, Dr. Pravin Kshirsagar

Computer and Decision Making: An International Journal · 04.11.2026 · 3 (6S) · 3754-3764

DOI: 10.1922/comdem.v3i6s.19160 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computer and Decision Making: An International Journal, online erschienen am 04.11.2026. Verfasst von Firoz Akhtar und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Opportunities Melanoma is one of the most aggressive and potentially fatal forms of skin cancer, and its early diagnosis remains a significant clinical challenge due to its visual resemblance to benign skin lesions. This study aims to develop an accurate and reliable deep learning-based framework for automated melanoma detection to assist dermatologists in clinical decision-making and improve patient outcomes through early intervention. The proposed methodology enhances existing pre-trained convolutional neural network (CNN) architectures by integrating three additional convolutional layers into each base model, thereby improving feature extraction capabilities and classification performance. The modified models are trained and evaluated using dermoscopic skin lesion images to classify lesions as malignant or benign. Model performance is assessed using widely accepted evaluation metrics, including accuracy, sensitivity, and specificity. Experimental findings demonstrate that the enhanced architectures significantly outperform their original pre-trained counterparts. Among the evaluated models, the Modified DenseNet201 model achieves the best overall performance, attaining an accuracy of 96.24%, sensitivity of 94.27%, and specificity of 98.58%, indicating its effectiveness in identifying melanoma while minimizing false diagnoses. Despite these promising results, the study is limited by its dependence on a specific dermoscopic image dataset, which may restrict generalizability across different populations, imaging devices, and clinical environments. Therefore, further validation using larger, multi-center, and real-world datasets is recommended. The proposed framework offers practical benefits by providing rapid and accurate diagnostic support, reducing diagnostic errors, and decreasing the workload of healthcare professionals. Moreover, early and reliable melanoma detection can contribute to lower mortality rates and improved patient quality of life, particularly in regions with limited access to experienced dermatologists. This study provides a valuable contribution to medical image analysis through its enhanced deep learning architecture for skin cancer classification.

Zitieren

APA

Akhtar, F., Yadav, D. K. P., & Kshirsagar, D. P. (2026). Progenesis of Skin Disease Detection Using Deep Learning. Computer and Decision Making: An International Journal, 3(6S), 3754-3764. https://doi.org/10.1922/comdem.v3i6s.19160

Vancouver

Akhtar F, Yadav DKP, Kshirsagar DP. Progenesis of Skin Disease Detection Using Deep Learning. Computer and Decision Making: An International Journal. 2026;3(6S):3754-3764. doi:10.1922/comdem.v3i6s.19160

BibTeX
@article{Akhtar2026,
  author = {Akhtar, Firoz and Yadav, Dr. K. P. and Kshirsagar, Dr. Pravin},
  title = {Progenesis of Skin Disease Detection Using Deep Learning},
  journal = {Computer and Decision Making: An International Journal},
  year = {2026},
  volume = {3},
  number = {6S},
  pages = {3754--3764},
  doi = {10.1922/comdem.v3i6s.19160}
}
Kostenlos starten →