Published Paper Details:

Diabetes is a chronic condition. Traditional invasive techniques like the finger-prick test, though commonly used, often lead to discomfort and reduced patient compliance. To overcome these limitations, this project presents a Non-invasive monitoring system. The proposed system utilizes a Near-infrared (NIR) sensor by analyzing how light is absorbed and scattered through skin tissue instead of directly connecting the NIR sensor to the Raspberry Pi 4B. An ESP8266 Wi-Fi Module is used. The sensor data is sent to the Raspberry Pi 4B via serial communication. This setup ensures smooth data acquisition while maintaining modularity and flexibility in hardware design. The received data is processed using a Python-based long short-term memory (LSTM) neural network algorithm. The LSTM model is known for its ability to analyze time-series data and predict results accurately. In addition to glucose estimation, the system is also designed to predict other health parameters such as blood group, blood level, cholesterol level, skin temperature, hydration level and blood pressure using AI-based analysis.