Biological tissues identification from their spectral signals acquired by a Raman needle

In this paper, biological tissues are discriminated based on their intrinsic Raman spectral features. First, a Raman needle, which comprises of a Raman probe and a puncture needle, is devised to insert into biological tissues and acquire their Raman spectral data. The Savitzky-Golay filter is used to remove the data noise, and an adaptive iterative penalized least squares method to rectify the baseline. To extract the spectral features from the data, the principal component analysis (PCA) is used. Then, a genetic algorithm (GA) based support vector machine (SVM) method is proposed to classify the biological tissues based on the spectral features. Experimental study was conducted with different animal specimens and approved that the proposed methods can identify efficiently the different biological tissues.