Effectiveness Evaluation of Public Security Face Recognition Systems Based on Improved Unascertained C-Means

In order to evaluate and improve the effectiveness of the public security face recognition systems reasonably and scientifically in real scenes, we create an effectiveness evaluation index system and the improved Unascertained C-means model. The data of indicators are obtained through measuring practical application systems of five public security bureaus. Then we carry out data statistics and pretreatment. The improved Unascertained C-means algorithm is realized by programming in MATLAB. After inputting the data into the model, we obtain the evaluation results. By comparing the real achievements of each system during the evaluation cycle, we verify the consistency between the evaluation results and realities. On the basis of the comparative analysis of the evaluation results, we identify the weaknesses of a system, and put forward some optimization strategies of the system effectiveness. In the paper, we synthetically apply various methods, such as qualitative analysis, quantitative analysis, theoretical analysis and experimental analysis, to study the uncertainty problem of the effectiveness evaluation of the public security face recognition systems. The evaluation process is objective and impartial, the results are valid and reasonable, and the strategies are targeted and operable. This paper is of certain theoretical significance to the research on effectiveness evaluation, and it is of certain practical significance to effectiveness improvement of the practical application systems.