Method and apparatus for determining position of pupil

An apparatus for determining a position of a pupil, comprising at least one processor and at least one memory for storing a program unit;

• wherein the program unit is executed by the at least one processor and comprises:
• a first acquisition module, configured to acquire an image to be detected comprising a pupil;
• a second acquisition module, configured to acquire a binary image of a preset area based on a preset model of semi-supervised learning, wherein the preset area is an area where the pupil is located in the image to be detected;
• a third acquisition module, configured to acquire a centroid of the binary image;
• a determination module, configured to determine a central position of the pupil according to the centroid of the binary image;
• a fifth acquisition module, configured to acquire a first-type training set and a second-type training set, wherein the first-type training set and the second-type training set each comprises at least one original image to be trained;
• a sixth acquisition module, configured to acquire a network model, wherein the network model is configured to convert the at least one original image to be trained into the binary image;
• a first construction module, configured to construct a loss function of the network model;
• and
• a second construction module, configured to construct the preset model according to the first-type training set, the second-type training set and the loss function of the network model.

A non-transitory storage medium storing programs that, when executed, are used for executing a method for determining a position of a pupil, wherein the method comprises:

• acquiring an image to be detected comprising the pupil;
• acquiring a binary image of a preset area based on a preset model of semi-supervised learning, wherein the preset area is an area where the pupil is located in the image to be detected;
• acquiring a centroid of the binary image;
• determining a central position of the pupil according to the centroid of the binary image;
• wherein before the acquiring a binary image of a preset area based on a preset model of semi-supervised learning, the method further comprises:
• acquiring a first-type training set and a second-type training set, wherein the first-type training set and the second-type training set each comprises at least one original image to be trained;
• acquiring a network model, wherein the network model is configured to convert the at least one original image to be trained into the binary image;
• constructing a loss function of the network model;
• and
• constructing the preset model according to the first-type training set, the second-type training set and the loss function of the network model.