Several efforts have been made in the recent decades to achieve notable accuracy in iris recognition systems which is proved as reliable biometric for personal identification. Iris segmentation is the most critical part of iris recognition system a successful segmentation guarantees a successful recognition. Most of the iris based systems are designed for highly controlled environment so these designed models usually fail on non-cooperative environments. Several researches are made for better segmentation method to increase the overall system performance. In this paper significance of these methods will be discussed in detail with a focus on their applications in iris recognition systems.