Gait recognition method with monitoring mechanism

Abstract

Disclosed is a gait recognition method with a monitoring mechanism. The gait recognition method is characterized in that a three-layer dynamic Bayesian network model is adopted, a first layer of the model is featured by the contour of gait of a human body, and a state of the first layer at a current moment is only relevant to a state of the first layer at a next moment; a second layer of the model is featured by frame difference images of the gait, and a state of the second layer at the current moment is relevant to a state of the second layer at the next moment and is relevant to the state of the first layer at the current moment and the state of the first layer at the next moment; and a third layer of the model is a monitoring layer and is relevant to the state of the second layer at the current moment and the state of the first layer at the current moment. The gait recognition method has the advantages that the traditional model structure is improved, and the gait recognition method is additionally provided with the monitoring mechanism; optimization learning for parameters is obviously improved, and the gait recognition rate is greatly increased; and the gait recognition method has an actual application significance and enormous social benefit.

Description

technical field [0001] The invention relates to the technical fields of computer science and image processing, and in particular provides a gait recognition method with a supervision mechanism. Background technique [0002] As a remote biometric authentication technology, gait recognition has attracted more and more attention. Gait recognition is to carry out remote identity authentication according to the posture of people walking. Gait is non-invasive and difficult to camouflage. Unlike fingerprint or iris recognition, which requires close contact with the recognized target during feature extraction, gait is the most potential biological feature in the field of long-distance video surveillance. [0003] In recent years, the research on gait recognition has been increasing, most of which are identified by extracting the gait contour of people walking. The technology can be divided into two categories; one is to use the static information of the gait contour, such as Kim pr...

AI Technical Summary

Problems solved by technology

The second is to use the dynamic information of the extracted contour. There are many algorithms for this kind of research, such as the gait recognition method based on gait energy image (GEI) and 2-dimensional principal component analysis [Wang Kejun, Liu Lili, Ben Xunye. Based on gait energy Gait recognition method based on image and 2-dimensional principal component analysis [J], Chinese Journal of Image and Graphics, 2009, 14(12): 2503-2509], this method uses GEI image as gait feature image, and performs 2-dimensional principal component analysis, but also because the GEI image is only an image composed of the average value of human gait frames, it must be affected by coats, backpacks, etc., and has a higher recognition rate in the single-class training samples of normal, coat, and backpack. , The recognition rate is low in the normal hybrid mode, such as Figure 2shown; Figure 2 is the same person walking normally , wearing a jacket, and a GEI image under a backpack, it can be seen that the static information of the outline of the GEI image in the three states is very different, so the recognition rate is low; in order to improve the loss of too much dynamic information in the GEI image, Zhang proposed based on Active energy image plus 2DLPP for gait recognition [J], Signal Processing, 2010, 90(7): 2295-2302 ], where the AEI image is composed of superimposed frame difference images, which can well reflect the dynamic gait characteristics of the human body when it is moving, and the recognition method is novel, but at the same time, the AEI image ignores the static information of the human body, such as Figure 3, it can be seen that the AEI image expresses the dynamic information very well, especially It is the leg movement information, but there is also the influence of coats and backpacks on the static information of the contour; in order to preserve the static information of the gait walking, Chen proposed to establish the frame difference energy map (FDEI) for the extracted incomplete gait contour Construct gait features by means of the method, and at the same time establish a hidden Markov model (HMM) to express [Chen C H, Liang J M, Zhao H. Frame differ ence energy image for gait recognition with incomplete silhouettes [J], Pattern Recognition Letters, 2009, 30(11):977–984.], the recognition effect is good, and the frame difference energy map (FDEI) is superimposed with the frame difference image , and it is summed with the static part of the human body while walking to form an FDEI image. The FDEI image can well express the static and dynamic information of the human body; such as Figure 4 shows that the frame difference image can express the motion characteristics very well, but due to the immobile part of walking with gait As static information, the formed FDEI image is still greatly affected by the backpack

Images