Abstract: An action recognition method includes: obtaining original feature submaps of each of temporal frames on a plurality of convolutional channels by using a multi-channel convolutional layer; calculating, by using each of the temporal frames as a target temporal frame, motion information weights of the target temporal frame on the convolutional channels according to original feature submaps of the target temporal frame and original feature submaps of a next temporal frame, and obtaining motion information feature maps of the target temporal frame on the convolutional channels according to the motion information weights; performing temporal convolution on the motion information feature maps of the target temporal frame to obtain temporal motion feature maps of the target temporal frame; and recognizing an action type of a moving object in image data of the target temporal frame according to the temporal motion feature maps of the target temporal frame on the convolutional channels.

Abstract: A phase loss detection device, a compressor including the same, and a phase loss detection method are disclosed. The phase loss detection device may include a signal converting circuit and a processor. The signal converting circuit is configured to convert voltage signals corresponding to respective phases of multiphase alternating current (AC) power monitored from a motor. The processor is configured to receive the converted voltage signals from the signal converting circuit, and configured to calculate, based on the converted voltage signals, one or more phase angles between the respective voltage signals. The processor is configured to determine that phase loss occurs if any one or more of the calculated phase angles deviate from a nominal value of a corresponding phase angle of the multiphase AC power by a value higher than a predetermined threshold. The phase loss detection can be performed in a convenient, effective and reliable way.