Intelligent classification of cervical pre-cancerous cells based on the FTIR spectra

Inspired from the great potential of the Fourier-transform infrared (FTIR) spectroscopy as a screening tool for cervical cancer, this paper proposes an intelligent classification of cervical pre-cancerous cells based on the FTIR spectra. It consists of two parts: the extraction of FTIR characteristics and the intelligent classification of the pre-cancerous cells. Peak-corrected area-based features' extraction (PCABFE) is introduced as a tool for the first part, while the Hybrid Multilayered Perceptron (HMLP) network is employed to classify the cervical pre-cancerous cells according to the Bethesda classification. Correlation test proves the capability of the proposed PCABFE to be as effective as the manual extraction by human experts, while the HMLP network produces a good classification performance with 97.4% of accuracy.