Fingerprint template protection using multiline neighboring relation

• We have generated cancellable template for every minutiae in the fingerprint.
• Multiline neighboring relation shows the relation of reference minutiae and its neighbors.
• We have selected only neighboring minutiae which are present in the M rectangles drawn around every reference minutiae.
• Proposed method shows better performance in case of lost-token attack.
• Revocability and security analysis is analyzed.

Security and privacy concern of biometric information leads to the attention for generating cancellable template from the original biometric template. We proposed an alignment free cancellable template generation by construction of M rectangles with different orientations around every reference minutiae followed by the calculation of rotation invariant and translation invariant neighboring relation, plane based quantization for bit string generation and matching. The proposed method fulfills non-invertibility, accuracy, diversity and revocability properties which are the primary requirements of cancellable template design. The EER obtained for FVC 2002 DB1, DB2, and DB3 is 0.62%, 1.33%, and 2.64% for same key scenario. The proposed method shows the better performance for FVC 2002 DB1, DB2, and DB3 compared to the existing methods (Wang & Hu, 2012, 2014; Ahmad et al., 2011).