Identification of individuals using iris recognition is an emerging technology. Segmentation of the iris texture from an acquired digital image of the eye is not always accurate - the image contains noise elements such as skin, reflection and eyelashes that corrupt the iris region of interest. An accurate segmentation algorithm must localize and remove these noise components. Texture features are considered in this paper for describing iris and non-iris regions. These regions are classified using the Fisher linear discriminant and the iris region of interest is extracted. Four texture description methods are compared for segmenting iris texture using a region based pattern classification approach: Grey Level Co-occurrence Matrix (GLCM), Discrete Wavelet Transform (DWT), Gabor Filters (GABOR) and Markov Random Fields (MRF). These techniques are evaluated according to their true and false classifications for iris and non-iris pixels.
Reference:
Bachoo, A and Tapamo, J-R. 2009. Comparison of features response in texture-based iris segmentation. SAIEE Africa Research Journal, Vol. 100(1), pp 2-11
Bachoo, A., & Tapamo, J. (2009). Comparison of features response in texture-based iris segmentation. http://hdl.handle.net/10204/3668
Bachoo, A, and J-R Tapamo "Comparison of features response in texture-based iris segmentation." (2009) http://hdl.handle.net/10204/3668
Bachoo A, Tapamo J. Comparison of features response in texture-based iris segmentation. 2009; http://hdl.handle.net/10204/3668.