Research On Authentication Method Based On Fingerprint Recognition

Using fingerprints in biometrics to complete identity authentication has improved convenience and security.The development of the Internet of Things and 5G has caused the explosive growth of various terminal devices,accelerated the progress of biometric authentication technology,and has more application scenarios.However,the complexity of the communication environment makes personal data including user fingerprints face a huge risk of leakage,and privacy cannot be effectively guaranteed.Therefore,the importance of studying how to combine other emerging technologies to build a fingerprint authentication scheme that can protect private data is self-evident.Aiming at the problems in the existing scheme,this thesis mainly does the following research:(1)A fingerprint classification fusion algorithm based on dual input is proposed.The model is based on a convolutional neural network,with fingerprint grayscale images as the main input,directional map or absolute gradient map as a supplement,and two feature-level fusion methods,single-scale and multi-scale,are set up.Experiments show that the form of main and auxiliary collocation can provide richer texture information,and the fusion of different learned features can reduce the data requirements and computational complexity of the model,and it is more robust and improves the classification accuracy when fuzzy samples exist.(2)A cross-domain authentication scheme in blockchain based on fingerprint identification is proposed.The scheme designs two anonymous cross domain authentication parts,which take fingerprint and password as two factors,and make use of the distributed and tamperable characteristics of blockchain to write the auxiliary recovery value into the global account book,so as to solve the problems of insufficient protection of user identity and location information,and easy attack and usurpation of auxiliary recovery value.Analysis and comparison show that the scheme has more security properties,moderate computing cost,and can meet the needs of different scenarios.(3)A fingerprint based anonymous authentication scheme in telemedicine environment is proposed.The scheme uses fingerprint as authentication factor,uses fuzzy extraction technology to process fingerprint template to avoid leakage,and uses public-private key encryption and session key agreement to protect the privacy and security of personal medical report and other data.Analysis and comparison show that the scheme can resist replay and other attacks,ensure the authenticity of the message and confidentiality of medical data,has high efficiency of single round authentication,and can provide privacy protection anonymous authentication service.