Fig. 1. (a) Schematic diagram of energy level structures for Rayleigh scattering and Raman scattering processes; (b) schematic diagram of Rayleigh scattering and Raman scattering spectra

Fig. 2. (a) Schematic diagram of a handheld Raman spectroscopy detection system; (b) schematic diagram of confocal Raman microscopy system

Fig. 3. Localized surface plasmon resonance^[13]

Fig. 4. Electric field enhancement mechanism of SERS^[13]

Fig. 5. Chemical enhancement mechanism of SERS, modified from reference [16]. (a) The non resonant enhancement between the tested molecule and metal nanoparticles is independent of excitation; (b) resonance enhancement formed by laser energy and electronic transitions within the tested molecule; (c) class resonance enhancement of photo induced charge transfer

Fig. 6. Precious metal nanoparticles with different geometric structures. (a) Spherical^[25]; (b) rod-shaped^[26]; (c) triangular^[27]; (d) star-shaped^[28]; (e) cage^[29]; (f) core-shell cage^[30]

Fig. 7. General process of Raman spectroscopy data processing

Fig. 8. Workflow of machine learning^[54]

Fig. 9. Three common neural network models. (a) Multilayer perceptrons, convolutional neural networks, and deep tensor neural networks^[60]; (b) composition of convolutional neural network^[61]

Fig. 10. The application of incoherent Raman microscopy spectroscopy technology at the cellular level. (a) H&E staining maps, protein and lipid distribution maps, as well as spontaneous Raman spectra and classical least squares fitting of normal colon tissue (upper) and cancerous colon tissue (lower)^[63]; (b) SERS immunoassay principle diagram for epithelial mesenchymal transition^[64]; (c) a blood biochemical map drawn at the cellular level using spontaneous Raman microscopy spectroscopy^[65]; (d) mapping metabolic changes in endothelial cells using spontaneous Raman probes^[66]

Fig. 11. Application of Raman microscopy spectroscopy technology at the tissue level. (a) SERS spectrum, SERS imaging, and photoacoustic imaging images of tumor tissue in breast cancer mouse model^[74]; (b) single and multiple SERS imaging detection of coronary artery endothelial cell tissue^[75]

Fig. 12. The application of Raman microspectral technology in body fluid biopsy. (a) Schematic diagram of SERS technology combined with artificial intelligence using exosomes in plasma to detect various cancers^[89]; (b) schematic diagram of the preparation process of saliva protein silver nanoparticles mixture, and comparison of SERS spectra of the mixture with saliva protein and silver nanoparticles without silver^[90]

Fig. 13. Application of incoherent Raman microscopy in microbiology. (a) Schematic diagram of identifying bacteria from Raman spectra using CNN algorithm^[112]; (b) schematic diagram of SERS technology combined with machine learning algorithms for classifying viruses in saliva^[113]