DLLE model-based data dimension reduction and characteristic understanding method

Abstract

The invention discloses a DLLE (Linear Local Embedding of Difference) model-based data dimension reduction and characteristic understanding method, and belongs to the field of computer vision. The method comprises the steps of firstly, obtaining an image sequence through a visual sensor, then analyzing an input motion image sequence, extracting a foreground human body contour region through a background subtraction method, performing binarization, researching a periodic characteristic of a motion, performing key frame extraction on each motion sequence, and extracting a complete motion periodic sequence; performing manifold dimension reduction through a DLLE algorithm to obtain a low-dimensional eigenvector, and storing the low-dimensional eigenvector in a motion database; and performing identification through a nearest neighbor classifier by comparing a mean Hausdorff distance between a test sequence and a motion sequence in a training sample library. According to the method, the application of a differential function and category information-based neighborhood preserving embedding algorithm to human body motion identification is proposed; a DLLE model can not only keep a manifold local geometric structure during dimension reduction but also fully utilize category information of original high-dimensional data; and the extension from unsupervised extension to supervised extension is realized.

Description

technical field [0001] The invention belongs to the field of computer vision, and proposes a new selective intelligent dimension reduction method based on a difference function to improve the accuracy of human action recognition. Background technique [0002] In the field of computer vision, human action recognition technology is a relatively important research topic. Most of the objective world information needed by human beings in daily life is obtained through vision. With the advancement of science and technology, it has become a new trend to make computer vision have visual effects similar to human beings. Not only does an intelligent computer have the ability to learn, understand, and analyze autonomously, but it also requires that it can recognize and understand the actions of targets in the external environment and perform higher-level analysis. At present, the methods of human action recognition are mainly divided into methods based on template matching and methods...

AI Technical Summary

Problems solved by technology

The sparseness of sample point density and the correlation between data sets are problems that the traditional LLE algorithm does not solve well in the process of extracting low-dimensional features of images. In the end, it often leads to mapping from high-dimensional to low-dimensional, and the obtained low-dimensional human body flow There is a certain deviation in the shape structure vector, which affects the recognition effect and robustness

Images