A Human Behavior Recognition Method Based on Random Projection and Fisher Vector

Abstract

The invention discloses a character behavior recognition method based on random projection and Fisher vector, which adopts the method of random projection theorem to replace the principal component analysis method for feature dimensionality reduction, so as to solve the problems of large time consumption and unclear retention of principal components. The projection theorem shows that through a compressed measurement matrix, the original signal with sparse properties can be projected onto a low-dimensional subspace, and the point distance between the mapped vector and the original high-dimensional feature vector remains basically unchanged, that is, the entire compressed The process does not produce a misinterpretation of the data. In addition, different from the hard division of the BoW model, the present invention adopts the GMM-Fisher vector hybrid model to softly divide the trajectory feature vector, which combines the characteristics of the Fisher kernel generation mode and the discrimination mode, and can not only calculate the occurrence of each characteristic descriptor Frequency can also describe the probability distribution of these feature descriptors in a statistical sense, which not only enriches the feature expression of behaviors, but also improves the efficiency of behavior recognition.

Description

technical field [0001] The invention relates to the technical field of signal processing, in particular to a character behavior recognition method based on random projection and Fisher vector. Background technique [0002] Behavior recognition technology is widely used in video surveillance, video retrieval, military detection, medical diagnosis and monitoring, and has broad application prospects and economic value. The traditional behavior recognition method is to embed the extracted trajectory features into the visual bag-of-words (Bag-of-Words, BoW) model, construct a rich visual dictionary by extracting local trajectory features in the video, and use the central clustering The frequency of the local feature vector relative to the central word is counted by the method, and the histogram formed by the visual word frequency is used to represent a type of video and finally achieve the purpose of character behavior recognition. The key to the BoW model is to construct a very...

AI Technical Summary

Problems solved by technology

However, this method cannot handle data on nonlinear manifolds, and the calculation process is quite time-consuming, requiring a large amount of storage space, which seriously affects the efficiency of classifier training and classification.

Images