Method and device for classifying videos

The embodiment of the invention discloses a method and a device for classifying videos, relates to the technical field of electronic information, and aims at improving the accuracy of video classification. The method comprises the steps of sectioning videos in a sample video database according to a time sequence to obtain a sectioning result; generating a moving atom set; generating a moving phrase set which can express a complex moving mode according to the moving atom set and the sectioning result; generating a description vector of the videos in the sample video database based on the moving phrase set; determining the video which has the same type with video in the sample video database and is to be detected according to the description vector. The method is applicable to video classification scenes.

[1]  Limin Wang,et al.  Motionlets: Mid-level 3D Parts for Human Motion Recognition , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[2]  Venu Govindaraju,et al.  Language-motivated approaches to action recognition , 2013, J. Mach. Learn. Res..