Scene recognition method and device, electronic equipment and readable storage medium

Abstract

The embodiment of the invention provides a scene recognition method and device, electronic equipment and a readable storage medium. The method comprises the steps: obtaining a to-be-processed image, and carrying out the recognition of the to-be-processed image to obtain a target person; detecting the target person to obtain human body posture information for the target person; encoding the human body posture information to obtain a human body posture encoding vector; performing regression prediction on the human body posture coding vector to generate a scene score for the to-be-processed image; wherein the scene score comprises a discretized scene score; and recognizing the scene of the to-be-processed image according to the discretized scene score. According to the embodiment of the invention, based on the scene score of the human body posture of the target person, the scene characteristics of the to-be-processed image are finely measured, and the scene of the to-be-processed image can be visually recognized with any granularity under the condition of discretizing the scene score.

Description

technical field [0001] The present invention relates to the technical field of image recognition, in particular to a scene recognition method, a scene recognition device, an electronic device and a computer-readable storage medium. Background technique [0002] The scene difference refers to the difference in the size of the range that the subject appears in the camera viewfinder due to the different distances between the camera and the subject. From near to far, it can be roughly divided into close-up, close-up, medium shot, panorama, and distant view. Alternately using various scenes in video works can make the narrative of the film’s plot, the expression of the characters’ thoughts and feelings, and the handling of the relationship between the characters more expressive, thereby enhancing the artistic appeal of the film. Effectively identifying the scene types of video images, on the one hand, helps to improve the efficiency of material retrieval in the later stage, and ...

AI Technical Summary

Problems solved by technology

[0003] Most of the existing methods directly classify the image to identify the scene of the picture. On the one hand, this classification method is limited by the number of categories, and can only be recognized at a coarser granularity; on the other hand, due to the definition of the scene Certain requirements, for example, define the scene above the chest as the close shot, and define the scene above the waist as the middle ground, for the ambiguous scene, that is, for the scene between the chest and the waist It is difficult to accurately identify its characteristics; at the same time, this classification method has higher requirements for data collection

Images