Protection of biometric data has assumed critical importance in the current biometric community. One way of doing this, is cancelable biometric method, which deals with the transforming the biometric data before storing in the database, in such a way that relative minutiae information in the transformed template is not degraded. This report contains a study and implementation results of some methods of generating cancelable biometric templates. The report contains the implementation of the method presented in the paper by Lee et al.  and development and implementation of two other methods. In order to develop a new method for biometric template transformation, two different methods have been implemented and their performance analysed. In the first method, minutiae similarity score is calculated to preserve the relative geometric minutiae information without requiring for pre-alignment of fingerprint templates. One minutia is selected as the reference minutia and all the other minutiae are rotated and translated around the reference minutia. To calculate the similarity score for a minutia we measure its normalised distance from the reference minutia, the cosine of the difference in the angle of this minutia and the reference minutia and the type of the two minutiae. A “type score” of 1 is generated if the both the minutiae are of the same type; otherwise a score of -1 is generated. The cosine of the difference in the angle and the normalised distance are added and the sum is multiplied with the “type score” to generate the entries in the template. The order of the string template is permuted according to the type of the reference minutia and user’s PIN. Finally, cancelable templates are generated by changing the reference minutia in turn. The second method involves the determination of the average distance of all the minutiae from the reference minutia. Then a line segment, whose length is equal to the average distance, is drawn in the direction of orientation of the reference minutia. This line segment is divided (quantised) into a number of parts, which depends on the quantisation number. Next, perpendiculars from all the minutiae are dropped onto this line segment and the number of foot of perpendiculars are summed in each “bin”. These numbers form the entries in the template. A template is generated in this way and its order is permuted depending on the user’s PIN. In the implementation of the first method, results have been analysed using the database FVC2004DB2. The results for the second method have been analysed using the same database but instead of using the complete database only a subset was used due to the lack of time.