Biometric user authentication has grown significantly in the previous decade. The primary challenge with the usage of biometric technology is the privacy and security of the biometric data. The literature suggests the use of cancelable biometrics to overcome these difficulties. This paper proposes three simple but efficient methods to generate cancelable biometric templates. The proposed methods combine non-invertibility and biometric salting. The methods are based on random permutation and bitwise-XOR. The first method combines the original biometric image with a random matrix using addition. The second method uses subtraction for salting. The cancelable biometric templates are generated using bitwise-XOR in both of these methods. The third method uses bitwise-XOR directly for template generation without salting. The methods are evaluated using Georgia Tech face, UBIRIS.v1 iris, and IITD ear biometric image datasets. The resulting templates’ quality is evaluated using several criteria outlined in existing scholarly literature. The proposed method’s ease of implementation and efficacy are its key advantages.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Bitwise-XOR for Cancelable Biometric

  • Onkar Singh,
  • Ajay Jaiswal,
  • Naveen Kumar

摘要

Biometric user authentication has grown significantly in the previous decade. The primary challenge with the usage of biometric technology is the privacy and security of the biometric data. The literature suggests the use of cancelable biometrics to overcome these difficulties. This paper proposes three simple but efficient methods to generate cancelable biometric templates. The proposed methods combine non-invertibility and biometric salting. The methods are based on random permutation and bitwise-XOR. The first method combines the original biometric image with a random matrix using addition. The second method uses subtraction for salting. The cancelable biometric templates are generated using bitwise-XOR in both of these methods. The third method uses bitwise-XOR directly for template generation without salting. The methods are evaluated using Georgia Tech face, UBIRIS.v1 iris, and IITD ear biometric image datasets. The resulting templates’ quality is evaluated using several criteria outlined in existing scholarly literature. The proposed method’s ease of implementation and efficacy are its key advantages.