Table 2.

Comparison of various non-invasive optical glucose sensing techniques [14,23,27,30,45,46,62,63].

Technology	Wavelength	Selectivity	Measurement Site	Merits	Drawbacks
NIR spectroscopy	750–2500 nm	Good	Ear lobe, finger, forearm, cheek, lip mucosa, oral mucosa, and tongue	-Low-cost -Easy to implement	-Glucose heterogeneous distributions affect accuracy. -Interferences by other chemical compounds
MIR spectroscopy	2500–10,000 nm	Good, superior to NIR	Finger, skin, and oral mucosa	-Quite accurate -Lightweight -Scattering is low	-Poor skin penetration depth -Expensive -High water abortion
FIR spectroscopy	10–1000 μm	Good	ISF	-Scattering is lower than NIR and MIR -Individual daily calibration is not required	-Difficulty in identifying other molecules than water due to strong water absorption
Raman spectroscopy	Visible light	Excellent	Eye, human skin	-Low sensitivity to water and temperature changes -Great specificity -Low-cost	-Lack of stability in the laser wavelength and intensity -Spectrum acquisition takes time
THz-TDS	30 µm to 3 mm	Good	ISF	-Not affected by background noise.	-Long measuring time -Low spatial and depth resolution
Fluorescence	Ultraviolet light, visible light	Excellent	Tears, human skin	-High sensitivity and specificity to glucose concentration -Not affected by light scattering	-Sensitive to changes in pH and oxygen levels -Susceptible to toxicity problems
PA spectroscopy	Ultraviolet light, NIR, and MIR	Good	Finger, forearm, and aqueous humor	-Unsusceptible to water distortion -Not affected by scattered particles	-Low signal-to-noise ratio -Affected by temperature changes, motion, pulsation, and acoustic noise