Hello Sashitra Devi,

Please go through this data herewith i have added it may help to you i think so,

In iris recognition systems, input samples are iris images. In this context, a noise source can be defined as anything that modifies the input image in such a way that the conditions for recognition are not optimal. In the case of iris, there are some noise sources associated to the existing conditions when the image is obtained. This way, blinking, a slight gaze deviation in the moment of capturing the image or an iris anomaly due to surgery or physiological reasons, are noise sources that make recognition difficult or even impossible. Another kind of noise sources, such glasses or contact lenses, are tangible. Five different groups of noisy images have been analysed.

In the first three groups, noise is provoked by non-tangible causes, while in the last two groups, tangible elements like glasses or lenses are needed.

• Gaze deviation: When acquiring the image, if not controlled, it is possible that the gaze is deviated to any direction. Depending on how off-axis the iris is, it can be directly processed or not. Gaze deviation can be corrected by using trigonometry to estimate the gaze angle and make a transformation to rotate the eye until it looks straight ahead.

• Eyelid obstruction: One of the most frequent noise sources in iris images is eyelid and eyelashes obstruction due to blinking. In this case, a logical mask must be set to prevent the non-valid pixels from affecting the iris-code. Also the amount of iris occluded must be taken into account, in order not to consider too little valid information.

• Mydriasis: Mydriasis is an excessive dilation of the pupil due to disease, trauma or the use of drugs or alcohol. Mydriasis can be also artificially provoked by using a mydriatic in the form of eye drops. In the case of mydriasis, non-elastic deformations of the iris occur as the pupil dilates. Due to this kind of deformations, the iris circular shape is affected. If mydriasis is in an advanced stage, recognition systems performance could be affected.

• Glasses: Considering that a big percentage of the population wears glasses, it is important that this noise source does not affect the iris recognition systems performance. Reflections are one of the adverse effects of glasses. If these reflections cover part of the iris, a logical mask must be set to prevent the non-valid pixels from affecting the iris-code, just as occurs with eyelids and eyelashes.

• Contact lenses: Due to certain pros of contact lenses, its use has become widespread. Depending on the kind of lens, iris recognition systems can be influenced in more or less extent. Conventional contact lenses are generally categorized as hard or soft lenses.

svm needed to be train by the features that extracted from train data (iris image) before test svm