In an embodiment, the face liveness recognition method further includes: obtaining an image sample set, where the image sample set includes a live facial image and a non-live facial image; obtaining a facial image in a corresponding image sample along a facial region of each image sample in the image sample set to obtain a first training sample; and training the first recognition model according to the first training sample.

The image sample set includes several image samples. The image samples may be live facial images and non-live facial images. The ratio of the number of live facial images to the number of non-live facial images may be 1:1 or other ratios.

Specifically, the server may obtain a facial image from the image samples in the image sample set to obtain a first training sample. The server may use a facial image obtained from a live facial image as a positive training sample, and use a facial image obtained from a non-live facial image as a negative training sample. Classification capabilities of the first recognition model are trained through the positive and negative training samples, so as to classify the target image as a live facial image or a non-live facial image.

In an embodiment, the training the first recognition model according to the first training sample includes: obtaining an initialized first recognition model; determining a first training label corresponding to the first training sample; inputting the first training sample into the first recognition model to obtain a first recognition result; and adjusting model parameters of the first recognition model according to a difference between the first recognition result and the first training label, and continuing training until satisfaction of a training stop condition.