Audio-Based Symptom Classification Using Deep Learning: The V-Cold Approach

Advanced deep learning methods were created to classify health symptoms from audio recordings and overcome drawbacks of conventional diagnostic methods, which frequently necessitated invasive procedures and specialized devices. Advanced acoustic characteristics such as RMS power, MFCC, and zero-crossing rate were used to accurately categorize symptoms. A CNN model was constructed to identify health symptoms from these features and demonstrated higher classification accuracy than existing methods. The V-Cold project provided a comprehensive evaluation of the model's performance using metrics such as precision, recall, Fl-score, and accuracy, highlighting the robustness and suitability of the method for diverse non-invasive diagnostic applications. Research explored the application of the V-Cold method in telemedicine and telehealth services, underscoring its potential to advance acoustic health diagnostics and expand access to immediate medical insights.