Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{198334,
        author = {K.LOKESWARI MALLIKA and M.BINDHU and R.JAYA AKHILESH and P.GANESH and DR.V.RAMESH BABU},
        title = {SKIN DISEASE DETECTION USING CONVOLUTIONAL NUERAL NETWORK},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {8611-8615},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=198334},
        abstract = {Dermatological conditions represent a significant portion of global healthcare challenges, often characterized by high diagnostic complexity and a shortage of specialized practitioners in rural or underserved regions. Traditional clinical diagnosis relies heavily on visual inspection, which is prone to human error and subjective interpretation. This research proposes an automated, high-precision framework for the classification of various skin diseases including Melanoma, Basal Cell Carcinoma, and Psoriasis utilizing Convolutional Neural Networks (CNN). The proposed model leverages the HAM10000 dataset, comprising thousands of multi-source dermatoscopic images, to train a deep learning architecture optimized for feature extraction and pattern recognition.
The methodology involves multi-stage image preprocessing, including noise reduction, hair removal filters, and data augmentation techniques to address class imbalance. We implemented a custom CNN architecture alongside transfer learning models such as ResNet-50 and MobileNetV2 to evaluate performance benchmarks. Our results demonstrate a significant improvement in diagnostic accuracy, achieving an overall sensitivity of over 94% and a specificity of 92%, outperforming standard baseline models. Furthermore, the integration of a Softmax classification layer allows for the probabilistic identification of malignant versus benign lesions, providing a critical tool for early-stage screening. This study concludes that CNN-based computer-aided diagnosis (CAD) systems offer a scalable, cost-effective, and highly reliable solution for enhancing clinical decision-making, potentially reducing the mortality rates associated with late-stage skin cancer detection.},
        keywords = {Convolutional Neural Networks (CNN), Deep Learning, Dermatology, Skin Lesion Classification, Computer-Aided Diagnosis (CAD), ISIC Dataset.},
        month = {April},
        }