Method and device for predicting intelligent man-machine interaction system with body based on action recognition

Abstract

The invention relates to the technical field of man-machine interaction, in particular to a method and a device for predicting an intelligent man-machine interaction system with a human body based on action recognition, and the method comprises the steps: collecting three-dimensional coordinates of a plurality of skeleton key points of a human body and time frames corresponding to the coordinates of each skeleton joint point, so as to generate initial skeleton data, the method comprises the steps of acquiring skeleton data, preprocessing the skeleton data to obtain final skeleton data meeting conditions, performing skeleton coding on the final skeleton data to extract at least one skeleton feature, and training a prediction model based on the at least one skeleton feature, the prediction model comprising a Kronecker lamination layer, a Kernel Matrix layer and a multi-layer perceptron layer to identify a future motion trajectory. Therefore, the problems that in the related technology, due to the fact that a large model is large in parameter quantity, high in data dependency and difficult to visually display the relation between variables, effectiveness and interpretability of the model in practical application are limited, and trust of a user to an identification result is affected are solved.

Images