Application of data fusion in computer facial recognition

Abstract

The recognition rate of single recognition method is inefficiency in computer facial recognition. We proposed a new confluent facial recognition method using data fusion technology, a variety of recognition algorithm are combined to form the fusion-based face recognition system to improve the recognition rate in many ways. Data fusion considers three levels of data fusion, feature level fusion and decision level fusion. And the data layer uses a simple weighted average algorithm, which is easy to implement. Artificial neural network algorithm was selected in feature layer and fuzzy reasoning algorithm was used in decision layer. Finally, we compared with the BP neural network algorithm in the MATLAB experimental platform. The result shows that the recognition rate has been greatly improved after adopting data fusion technology in computer facial recognition. DOI: http://dx.doi.org/10.11591/telkomnika.v11i11.3384