Health Intelligent Systems Improve Value of Cancer Care and Prognosis: A Proposed Medical System and Model For Disease Management and Detection

Abstract

Skin cancer has emerged as a prevalent form of cancer, witnessing an upward trend in incidence over recent decades. The traditional methods for skin cancer identification are time-consuming and resource-intensive. Presently, the field of medical science leverages digital technology tools for efficient skin cancer classification. This study addresses challenges stemming from a shortage of annotated data samples for binary classification in skin cancer. Within this investigation, we introduce the single convolutional neural network (S-CNN) with multi-output functionality. The architecture of the S-CNN is intricately designed, encompassing multiple layers dedicated to extracting low to high-level features from skin images. Additionally, we integrate customized transfer learning models, specifically VGG-16 and VGG-19, into our study. Experiments were conducted utilizing a dataset comprising benign and malignant cases. The S-CNN model showcased remarkable accuracy, achieving a 96.66% success rate in effectively distinguishing between benign and malignant instances. Our automated model consistently demonstrated exceptional accuracy and performance in a comprehensive comparison with alternative methodologies.

Ähnliche Arbeiten