Graph based Cross-shape Recognition for Palm Diagnosis

According to the palm medicine (PM), the palmprint is one of the biometric features which can be used to diagnose diseases. Some special shapes formed by the palm lines in some areas of the palm indicate some special diseases. Cross-shape ("+") is very important in PM. This paper proposes a novel approach to identify the cross-shape for palm diagnosis. In this approach, the palm lines are extracted firstly, and then an undirected graph (called palm line undirected graph, PLUG) is constructed to represent the structure of the palm lines. Finally, by using the theory of fuzzy set, the cross-shapes are recognized in the PLUG. The experimental results demonstrate the effective of the proposed approach