Rearch And System Design Of Identity Recognition Fusion Algorithm Based On ECG And PPG Signals

With the rapid development of modern science and information technology and the widespread application of big data,these advances have provided convenience for modern people’s life,work and other scenarios,and the major issues of network information security that have come with it have also been valued by people.At present,identification technology based on biometrics is a hot topic of research.Due to the large differences in the physiological signals between different individuals,it has the advantages of good confidentiality and uniqueness,so it can be used for human identity verification.In this paper,the ECG signal and PPG signal in medical physiological signals are used as carriers to design and verify a set of networked identification systems based on multiple physiological signals.Firstly,the acquired ECG and the PPG signal preprocessing.Because the time-frequency domain features obtained by traditional feature extraction methods cannot fully represent the signal,the ideal recognition effect cannot be achieved.Therefore,in terms of PPG signal feature extraction,this paper proposes an improved orthogonal matching pursuit algorithm based on particle swarms.The best set of atoms selected in the super-complete atom library can reconstruct the original signal well,not only completes the sparse decomposition of the PPG signal,but also improves the search speed of the best atom,and takes the time-frequency characteristic parameters of the best matching atom as the characteristic of the PPG signal,a more ideal recognition effect is obtained,and the efficiency of the algorithm is greatly improved.Then,in order to deal with the situation that one of the individual signals may have a poor waveform while the other signal has a good waveform in actual applications,in order to enhance the stability of the system,this paper proposes a feature-level fusion algorithm of multi-modal signals based on canonical correlation analysis,which realizes feature complementation and weakens the influence of inherent defects of a single feature.In order to further shorten the running time of the algorithm and improve the performance of the system,this paper first uses the principal component analysis algorithm to achieve the first dimensionality reduction of the features,and then uses the CCA algorithm to perform the second fusion of the features to obtain the final fusion feature vector.Then built an identity recognition system model based on deep belief network,and imported the fusion feature vector into the model for training and recognition.Experiments show that the identification model based on PCA-CCA and deep confidence network has better classification and recognition effect,and the correct recognition rate is as high as 97.43%.Finally,this paper constructs a set of networked identification system based on ECG and PPG signal fusion algorithm.The hardware design of the system mainly has the functions of signal acquisition and network communication,while the software part is responsible for data communication,algorithm realization and visual interface display.Finally,experiments on the system verify the feasibility of the identification system proposed in this article.