The iris pattern is ideal as a biometric factor because it provides higher safety and accuracy than fingerprints and faces, but it is not widely used. The causes of existing iris recognition systems are the difficulty of introduction due to the use of large and complicated devices, and the low usability that input images must be captured at a close range. Accurate iris region segmentation is essential to describe individual features from captured images. However, if we aim to eliminate the constraint of the shooting distance while using small and simple imaging devices, it becomes difficult to segment the iris region because the absolute brightness of the pupil, which is the premise of existing methods, is unstable. Therefore, in this paper, we propose a method that enables stable iris region segmentation even in such images. We conducted evaluations combining the proposed method with typical feature extraction and matching methods. The results show that the proposed method achieves stable iris region segmentation in almost all images acquired in the target environment, and that has better overall certification performance than the typical existing method. It was also shown that highly accurate certification is available even when the contour of the iris region is ambiguous by using a novelly proposed method for noise removal together.