Gait behavior recognition method based on feature combination

Abstract

The invention discloses a gait behavior recognition method based on feature combination. The gait behavior recognition method based on feature combination includes the following steps that motion acceleration information of the body when a user behaves is obtained through an acceleration sensor; the peak value, frequency, gait cycle and interquartile range of each axis and the cross correlation coefficient between the different axes are calculated from the motion acceleration information; parameters are selected according to an aggregation method to form a feature vector; a sample set and the feature vectors of a gait acceleration signal to be identified serve as a training set, and a classifier is trained to have the capacity of classifying gait behaviors; all the feature vectors of the gait acceleration signal to be recognized are input to the trained classifier and given with fit categories, the fit categories of all the feature vectors are counted, and the category with the highest number of appearing times is given with the gait acceleration signal to be identified. The purposes that the calculation process is simplified, the number of dimensions of the feature vectors is decreased and the effectiveness is good are achieved.

Description

technical field [0001] The present invention relates to the field of body area network (Body Area Network) application technology, in particular to a gait behavior recognition method based on feature combination. Background technique [0002] At present, gait recognition has become a subject studied by more and more scientific researchers, and has become a new biometric authentication technology, which identifies a person's identity through the way people walk. Compared with some mature biometric methods, such as fingerprint recognition, face recognition and speech recognition, the advantage of gait recognition is that in addition to being unique, it is also non-contact, unaffected by distance, and difficult to detect. Camouflage, less affected by the environment and other characteristics. The uniqueness of gait features has important applications in many aspects, such as providing effective clues for human motion recognition and identification, and can play an effective au...

AI Technical Summary

Problems solved by technology

[0006] However, there are still some problems: (1) There are fewer samples and behavior types; (2) Multiple transformations are required, and the extracted feature dimension is too high, and dimensionality must be reduced; (3) Multiple complex calculations are required and optimized; ( 4) The design of the data acquisition system is relatively complicated

Images