FIELD: data processing.
SUBSTANCE: invention relates to video categorization facilities. Keyframe is obtained, which contains the face on the video. Facial features are obtained in the keyframe. One or more facial features that correspond to one or more image categories are obtained; in accordance with the facial feature in the keyframe and one or more facial features corresponding to one or more categories of images are obtained. Category of the images, to which this video belongs, is determined. Video is assigned to the category of the image, to which the video belongs.
EFFECT: improved accuracy of video categorization.
11 cl, 9 dwg
