Upon obtaining the video streams from the plurality of views of the game desktop, the video streams are input into the trained neural network for target object detection, so that the neural network for target object detection outputs the detection result in real time. The detection result indicates whether the specific target object is included in each image frame of the view streams. In the embodiments of the present disclosure, the target object may include, but not limited to, at least one of: a person, a game chip, a game card, a game currency, or the like. The specific target object may include a game card.

In the embodiments, the neural network for target object detection is used to detect whether the specific target object is included in each image frame of the video stream, so as to perform subsequent switching the game progress stage.

In some optional embodiments, the specific target object includes a game card, and step 103 includes: switching the game progress stage from a betting stage to a gaming stage at the time point corresponding to the first image frame in the plurality of the continuous image frames where the game card is detected.

In the embodiments of the present disclosure, if the game card is detected in the plurality of the continuous image frames in the video stream, it indicates that the present round of the game enters the gaming stage. Therefore, the time point corresponding to the first image frame in the plurality of the continuous image frames where the game card is detected is the time point for switching the game progress stage from the betting stage to the gaming stage.