Identity information processing method and device based on fundus image

A device of identity information comparison based on a fundus image, the device comprising:

• at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores an instruction executable by one processor, which when executed by the at least one processor, causes the at least one processor to perform operations comprising:
• recognizing a fundus image by using a neural network to obtain a multi-dimensional feature vector representing the identity of a user;
• comparing the obtained multi-dimensional feature vector with each pre-stored multi-dimensional feature vector in a database;
• and
• determining, according to the comparison result, whether there is a matching between a currently obtained multi-dimensional feature vector and a multi-dimensional feature vector pre-stored in the database;
• wherein the neural network is trained using triple sample data, wherein the triple sample data comprise a first fundus image sample, a second fundus image sample, and a third fundus image sample, the second fundus image sample and the first fundus image sample are fundus images of a same person, and the third fundus image sample and the first fundus image sample are fundus images of different people;
• wherein in a process of training the neural network, the neural network separately extracts multi-dimensional feature vectors of the first fundus image sample, the second fundus image sample and the third fundus image sample, calculates a first distance between the second fundus image sample and the first fundus image sample and a second distance between the third fundus image sample and the first fundus image sample according to the three extracted multi-dimensional feature vectors, obtains a loss value according to the first distance and the second distance, and adjusts parameters of the neural network according to the loss value;
• wherein the adjusting parameters of the neural network according to the loss value comprise:
• feeding back the loss value to the neural network, to enable the neural network to adjust the parameters according to the loss value to decrease the first distance and increase the second distance until the first distance is smaller than the second distance by a preset value.