Person and credential verification system and method based on deep learning

The invention relates to a person and credential verification system and method, in particular to a person and credential verification system and method based on deep learning and belongs to the field of security guarding. The system comprises a training subsystem and a person and credential verification subsystem; the method includes a training process and a person and credential verifying process; the training process includes adopting a plurality of on-site facial images of known identities and credential photos to train the training subsystem; the person and credential verifying process includes utilizing the module parameters acquired through the training process, comparing the plurality of on-site acquired images of a discharged person with the credential photos corresponding to the information acquired by a credential reading module, and outputting the verification result whether the person consistent with the credential or not automatically. By the aid of the system and method, the credential photos and a plurality of facial photos acquired through an on-site video monitoring device can be compared effectively, the adaptability to facial angles, on-site environment, light emission and other factors is high, the accuracy of person and credential verification is high, and the discharging speed is high.