The invention discloses an OCT
fingerprint identification method based on a three-dimensional
convolutional neural network, and the method comprises the steps: obtaining the OCT three-dimensional
fingerprint data of a reference
fingerprint and a to-be-detected fingerprint, and enabling the OCT three-dimensional fingerprint data of the reference fingerprint and the to-be-detected fingerprint to respectively comprise a
sweat gland and a
sweat gland; according to the OCT fingerprint identification method based on the three-dimensional
convolutional neural network, the three-dimensional data with the
sweat gland as the center is used as the input of the identification model, compared with a traditional two-dimensional epidermal
sweat pore identification method, feature information of a plane dimension is included,
depth dimension information is increased, three-dimensional morphological features such as a spiral structure of the sweat gland are completely presented, and the identification accuracy of the OCT fingerprint identification method based on the three-dimensional
convolutional neural network is improved. The multi-dimensional three-dimensional data can provide finer and more unique biological feature identification, and the discrimination and accuracy of fingerprint identification are effectively improved. Meanwhile, according to the method, identification is carried out based on three-dimensional data with sweat glands as the center, stable identification performance can be kept under severe conditions, the fingerprint anti-counterfeiting capability is further remarkably improved, and false fingerprint attacks are effectively resisted.