- Divakaran D
- Romaric Sallustre
Diabetes is one of the prominent diseases around the world. Presently, invasive techniques need a finger-prick blood sample. A repetitively painful procedure that produces the chance of infection. A device for blood Hemoglobin measurement was designed and developed without pricking the blood. This non-invasive device is based on principle of NIR spectroscopy technique with specific short wavelengths for detection of Hemoglobin molecule. The proposed technique is implemented using emitters and Camera of particular specifications and ratings. The data acquisition module is also implemented with combinations of emitters and Camera to collect the data. The proposed non-invasive device is examined by placing fingertip into the embedded clips which contain emitters and detectors. Hence, continuous blood Hemoglobin monitoring is possible through developed prototype. The mechanism of Hemoglobin detection is that when light of specific wavelengths passes through the object, then received resultant light is corresponds to change in blood Hemoglobin concentration. These collected spectrums are interpreted in the form of blood Hemoglobin concentration using a machine learning-based CNN model.
Index Terms — Non-Invasive Method, Blood Hemoglobin, Near-infrared (NIR) spectroscopy, CNN
- Python 3.8 or above.
- Pandas.
- Matplotlib.
- Numpy.
- CNN.
This project is under MIT licence.