We have developed an automated algorithm using hand-crafted features based on texture and color to identify abnormalities in the cervix. Combining multiple contrast mediators as inputs, in particular, contrast enhancement from acetic acid and lugol's iodine enables improved sensitivity and specificity compared to that using acetic acid alone.

The algorithm requires several seconds to analyze the images to provide a decision and when used with either the Pocket Colposcope or Callascope can render a rapid decision that is on par with the sensitivity of an expert provider but with superior specificity enabling more effective screening, triage and image guidance of treatment.