Fluorescence lifetime diagnosis of cervical cancer based on Extreme Learning Machine

Fluorescence Lifetime Imaging (FLIM) was used to study the histopathological conditions of cervical biopsy tissues. Measurements were conducted on more than 40 H&E stained cervical tissue sections. The characteristic decay lifetimes of the samples were extracted using an Expectation-Maximization and Bayesian Information Criterion algorithm. Diagnostic criterion based on the Extreme Learning Machine was developed to discriminate between normal and neoplastic samples. A high sensitivity and specificity of more than 80%were obtained. The proposed technique can be used to automate and supplement the traditional histopathological examination of cervical tissues.