Iris Pigment Spots Detection implementing Thresholding Method

Abstract

The increment size of the pigment spots on iris surface is indicated to the eye disease. Therefore, an automatic pigment spot detection has been proposed to detect the pigment spots on the iris surface. The main challenge is the type of feature that needs to be used for detection is unidentified. Based on the standard features applied for detection purposes, most of the features, such as shape, edges and vector, are not reliable. This situation occurs because the physical form of the pigment spots on the iris surface are dynamic. Hence, the pigment spots colour is the best feature possible to be applied, because it is moderately consistent. However, the spot colour intensity value are numerous. Several colour intensity values that have been used by other researchers were unable to detect the pigment spots. Henceforth, new colour intensity values based on thresholding method have been proposed in this paper. The approach has been applied through on the HSV colour model. The result shows the proposed values more accurate to detect the spots on the iris surface. The results have been recorded as follows (FAR) 0%, 1.33% and 4%, (FRR) 80%, 73.33%, and 70.67%, (DR) 20%, 25.33% and 25.33%.