Abstract: A dynamic facial expression recognition (FER) method based on a Dempster-Shafer (DS) theory improves a feature extraction effect of an expression video through multi-feature fusion, and deeply learns an imbalanced dynamic expression feature by using the DS theory, multi-branch convolution, and an attention mechanism. Compared with other methods, the dynamic FER method scientifically and effectively reduces an impact of sample imbalance on expression recognition, fully utilizes a spatio-temporal feature to mine potential semantic information of the video expression to perform expression classification, thereby improving reliability and accuracy and meeting a demand for the expression recognition.

Abstract: An unmanned device acquires sensing data of surrounding obstacles; determines, for each obstacle, at least one predicted track of the obstacle in a future period of time based on the sensing data; determines, for each moment in the future period of time and according to the predicted track corresponding to the obstacle, a collision probability that a collision with the obstacle occurs at each position in a target region at the moment; and determines a global collision probability that the collision with the obstacle occurs in the entire target region at the moment. According to the global collision probability corresponding to each obstacle at each moment, the unmanned device controls the unmanned device in the future period of time.