Palmprint based biometric identification has become an active research topic due to its richness in amount of features. In this paper, a combined approach for the segmentation of palm which is the region of interest. The acquired hand image is pre-processed and region of interest has been segmented where the more interesting features available. The segmentation can be performed by two alternative ways based on the finger positioning. Region based segmentation can be used in case of image with fingers spread apart. Centre key point based segmentation can be used in case of images with fingers positioned closely. Results of the proposed schemes ensure the reliability of the methods and reduce the constraint specified in the state of the art algorithms.