Sarcouncil Journal of Engineering and Computer Sciences

Abstract: This work describes an advanced machine learning method for Bowen's disease identification, which uses the VGG16 architecture. Bowen's disease is an early type of skin cancer that is defined by the proliferation of aberrant cells in the skin's outer layer. One major obstacle was the lack of publicly available datasets, which made gathering photos from various sources necessary. After ten training epochs, the model's astounding 99% accuracy was attained. Using more photos, post-training validation was carried out and the distinction between Bowen's illness and healthy skin states was effectively made. The promise of deep learning models in dermatological diagnoses is shown by this high accuracy. But compared to general skin cancer statistics, there are not many thorough records particularly for Bowen's disease, which emphasizes the need for more focused data gathering initiatives. According to this study, machine learning holds great potential for improving the precision of skin disease diagnosis.