Region of interest finding in reduced resolution colour imagery—application to cancer cell detection

摘要

This research explores image analysis and pattern recognition methods that eliminate the need to explicitly segment imagery prior to parameter measurement and object classification. The approach taken involves representation of key characteristics of objects to be recognized (or rejected) in terms of complex receptor fields and a flexible decision logic structure. A region of interest finding method for application in cancer cell detection in low-resolution three-colour “Pap” smear imagery implemented using this approach is described. Adaptations for use with colour imagery and the incorporation of parameter optimization techniques have resulted in greatly improved detection of rare “suspect” squamous cells in Pap smear imagery.