Action Recognition Method, Apparatus and Device, Storage Medium and Computer Program Product

An action recognition method performed by a computer device, comprising:

• obtaining multiple video frames in a target video;
• performing feature extraction on the multiple video frames respectively according to multiple dimensions to obtain multiple multi-channel feature patterns, each video frame corresponding to one multi-channel feature pattern, and each channel representing one dimension;
• determining an attention weight of each feature point based on a similarity between multiple feature points on multi-channel feature patterns of every two adjacent video frames, the attention weight being used for representing the degree of correlation between a corresponding feature point and an action performed by an object in the target video;
• and
• determining a type of the action based on the multiple multi-channel feature patterns and the determined multiple attention weights.

An action recognition apparatus, the apparatus comprising:

• an obtainer configured to obtain multiple video frames in a target video;
• a feature extractor configured to perform feature extraction on the multiple video frames respectively according to multiple dimensions to obtain multiple multi-channel feature patterns, each video frame corresponding to one multi-channel feature pattern, and each channel representing one dimension;
• an attention weight determiner configured to determine an attention weight of each multi-channel feature pattern based on a similarity between every two multi-channel feature patterns in the multiple multi-channel feature patterns, the attention weight being used for representing the degree of correlation between a corresponding multi-channel feature pattern and an action performed by an object in the target video;
• and
• a type determiner configured to determine a type of the action based on the multiple multi-channel feature patterns and the determined multiple attention weights.

An action recognition apparatus, the apparatus comprising:

• an obtainer configured to obtain multiple video frames in a target video;
• a feature extractor configured to perform feature extraction on the multiple video frames respectively according to multiple dimensions to obtain multiple multi-channel feature patterns, each video frame corresponding to one multi-channel feature pattern, and each channel representing one dimension;
• an attention weight determiner configured to determine an attention weight of each feature point based on a similarity between multiple feature points on multi-channel feature patterns of every two adjacent video frames, the attention weight being used for representing the degree of correlation between a corresponding feature point and an action performed by an object in the target video;
• and
• a type determiner configured to determine a type of the action based on the multiple multi-channel feature patterns and the determined multiple attention weights.