Raman spectroscopy proves to be a versatile technique for forensictasks. Whilst it is very easy and fast to record a spectrum, it may bevery time-consuming for the non-experts (and even the experts) toidentify chemicals from the large database of Raman spectra of thesematerials. Parallel developments in the field of neural networks havecome to a stage that they can participate well in the recognition ofthese materials. In this paper, a novel neural network-based techniqueis developed and presented for the quick identification of chemicalcompounds, in particular narcotics and explosives. Preliminary resultssuggest this hybrid network has a great potential for forensicapplications and for recognition of materials, in general
展开▼