Video grounding method and apparatus

A method of video grounding, comprising:

• obtaining video data and a query text;
• and
• using the video data and the query text as inputs of a trained video grounding model, and outputting information about a query video that is a video clip corresponding to the query text in the video data;
• wherein the trained video grounding model is used to extract features from the video data and the query text, respectively, to obtain a third video feature and a third text feature, fuse the third video feature and the third text feature according to an attention mechanism, use a fused feature as an input of an output layer, and output the information about the query video.

A video grounding apparatus, comprising:

• at least one processor;
• and
• at least one memory coupled to the at least one processor and storing a program, which when executed by the at least one processor, causes the video grounding apparatus to:
• obtain a video data set comprising video data;
• obtain a first video feature of the video data and a first text feature of the video data, wherein the first text feature comprises a plurality of words describing each frame of image in the video data;
• segment the first video feature, to obtain features of a plurality of video clips;
• map the plurality of words to the features of the plurality of video clips, to obtain a text description corresponding to each video clip of the plurality of video clips;
• and
• train a video grounding model based on the text description corresponding to each video clip, to obtain a trained video grounding model, wherein the video grounding model is used to output information about a corresponding video clip based on an input text.