Self-adaptive tiny action highlighting system

Abstract

The invention discloses a method for adaptively dividing a Fourier complex number space and carrying out tiny action highlighting. The method comprises the following steps: firstly, carrying out non-negative matrix factorization denoising and visual equalization preprocessing on a collected video; then, carrying out variational mode decomposition on different tiny action components existing in thevideo, splitting the video into sub-mode videos only containing a single action component according to a preset decomposition number, and then carrying out tiny action highlighting; and finally, filtering the video frame of each sub-modal video under different scales by the micro-motion highlighting processing module, extracting interested micro-motion frequency bands, performing phase amplification and inverse scale transformation, and performing superposition reconstruction on the same frame after inverse transformation of a plurality of modals to obtain a micro-motion highlighted video. Therefore, the effect of performing self-adaptive tiny action highlighting on the video is achieved, and a new tool is provided for the highlighting problem that various videos contain various tiny action components.

Description

technical field [0001] The invention relates to the field of video processing, and relates to a method for highlighting tiny movements. Background technique [0002] There are many small movements recorded by cameras in life. Although these tiny pixel-level movements are almost invisible to the human eye, they often contain important characteristics and intrinsic characteristics that reflect the object. For example, the periodic tiny blood vessel protrusions at the carotid artery are closely related to the blood supply to the human brain and can reflect vital signs; the tiny fluctuations caused by the breathing of creatures at the scene of natural disasters provide hope for the creatures to be rescued in the first place; bridges, The small vibration of the building under load greatly facilitates the analysis of its structural dynamic characteristics in a non-contact manner. In addition, the tiny movements of microorganisms, human cells, tissues, etc. that exist in the micr...

AI Technical Summary

Problems solved by technology

In the field of video processing, there have been some previous works on micro-motion highlighting technology, but there is still a bottleneck: the phase-based micro-motion highlighting technology proposed by predecessors (N Wadhwa et al.2013) is aimed at videos containing specific motion components Effective, but limited for videos containing many small motion components at the same time

Images