S-SCAN:AN ARTIFICAL INTELLIGENSE-BASED HYPERSPECTRAL SMARTPHONE PLATFORM FOR NON-INVASIVE MEDICAL

Abstract

Early detection of pathological conditions is a fundamental requirement for effective disease prevention and clinical intervention. However, conventional diagnostic procedures often rely on laboratory-based analyses, invasive sampling methods, and specialized medical equipment, which can limit accessibility and delay early diagnosis. This study introduces S-SCAN, an artificial intelligence (AI)–driven hyperspectral smartphone platform designed to enable non-invasive medical diagnostics through optical spectral analysis of biological tissues.The proposed system transforms a smartphone camera into a hyperspectral sensing device capable of capturing reflected optical signals from human tissues across multiple wavelengths. The collected spectral data are processed through a deep learning framework combining convolutional neural networks (CNN) for hyperspectral feature extraction and long short-term memory (LSTM) networks for temporal physiological signal modeling. This hybrid architecture enables the identification of complex spectral patterns associated with biochemical and metabolic changes in biological tissues.To ensure secure handling of sensitive medical information, the system incorporates blockchain-based data management and AES-256 encryption protocols, providing robust protection for patient data during storage and transmission. Experimental simulations and conceptual evaluations indicate that the integrated model can achieve diagnostic accuracy levels approaching 97–98%, while significantly reducing analysis time and resource requirements compared to conventional laboratory diagnostics.The results demonstrate that AI-driven hyperspectral mobile diagnostic systems have strong potential to support preventive healthcare, facilitate early disease detection, and expand access to medical services in remote or resource-limited environments.