Abstract: A method and apparatus for processing key points in an image are provided. The method includes: obtaining a first image region to which key points belongs from an (N-1)th frame image; obtaining a second image region to which the key points belong from an Nth frame image; analyzing optical flow of the key points between the first image region and the second image region, to determine position offset information of the key points; compensating first predicted position information of the key points in the (N-1)th frame image by using the position offset information to obtain compensated predicted position information of the key points; and performing temporal smoothing on second predicted position information of the key points in the Nth frame image using the compensated predicted position information to obtain smoothed predicted position information of the key points in the Nth frame image.

Abstract: A data processing method is provided. In the data processing method, target sequence data is obtained. The target sequence data includes N groups of data sorted in chronological order. Processing is performed, according to an ith group of data in the N groups of data, processing results of a target neural network model for the ith group of data, and a processing result of the target neural network model for a jth piece of data in an (i+1)th group of data, a (j+1)th piece of data in the (i+1)th group of data by using the target neural network model, to obtain a processing result of the target neural network model for the (j+1)th piece of data in the (i+1)th group of data, i being greater than or equal to 1 and less than N, and j being greater than or equal to 1 and less than Q.