On Combining Classifiers for Assessing Portrait Image Compliance with ICAO/ISO Standards

Abstract

The International Civil Aviation Organization (ICAO) has selected biomet- rics - in particular face - as the technique for travel documents to verify the association between such documents and the person in possession of them. The International Stan- dard Organization (ISO) has defined a standard for the digital face images to be used in Machine Readable Travel Documents. Due to the ISO/IEC 19794-5 international standard, there exists a high demand for automatically checking portrait images to as- sist civil service employees in decision making regarding ICAO/ISO compliance. We present a face normalization and analysis system implementing several requirements of the ISO/IEC 19794-5 specification. We address the criteria eyes-open and mouth- closed and highlight the fusion of complementary classifiers to boost performance of the overall analysis system. Our results show that classifier fusion is capable of im- proving the classification performance considerably as compared to a single classifier decision.