Optical spectroscopy has been shown to be an effective method for detecting neoplasia. Guided Therapeutics has
developed LightTouch, a non invasive device that uses a combination of reflectance and fluorescence spectroscopy for
identifying early cancer of the human cervix. The combination of the multispectral information from the two
spectroscopic modalities has been shown to be an effective method to screen for cervical cancer. There has however
been a relative paucity of work in identifying the individual spectral components that contribute to the measured
fluorescence and reflectance spectra. This work aims to identify the constituent source spectra and their concentrations.
We used non-negative matrix factorization (NNMF) numerical methods to decompose the mixed multispectral data into
the constituent spectra and their corresponding concentrations. NNMF is an iterative approach that factorizes the
measured data into non-negative factors. The factors are chosen to minimize the root-mean-squared residual error.
NNMF has shown promise for feature extraction and identification in the fields of text mining and spectral data analysis.
Since both the constituent source spectra and their corresponding concentrations are assumed to be non-negative by
nature NNMF is a reasonable approach to deconvolve the measured multispectral data. Supervised learning methods
were then used to determine which of the constituent spectra sources best predict the amount of neoplasia. The
constituent spectra sources found to best predict neoplasia were then compared with spectra of known biological
chromophores.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.