Design And Implementation Of Iris Recognition Evaluation System Based On Serverless Architecture

Iris recognition evaluation is an important part of algorithm quality assurance,and a reasonable and comprehensive evaluation can provide a comprehensive and accurate measure of the effectiveness of all aspects of iris recognition and the quality of the entire iris recognition software.Therefore,it is important to develop an automated iris recognition evaluation system that is easy to use and comprehensive.In this paper,we investigate the research related to iris recognition measurement at home and abroad and analyse the strengths and weaknesses of existing iris recognition measurements.Based on the findings,this paper presents the requirements analysis,overall architecture design and functional module design of the iris recognition assessment system.The key iris recognition assessment operations are analysed and the system is designed in terms of iris recognition assessment processes,interfaces,data and metrics.Simplifies the way in which iris recognition assessment can be performed by users without implementing a complete iris recognition system,and reduces the cost of assessment by providing algorithm files.To meet the requirements of concurrency,environmental isolation and scalability,this paper builds a serverless-based assessment cloud platform based on Knative.In the evaluation cloud platform,the message queue middleware is both a necessary decoupling middleware and a trigger source for evaluation cloud functions in the serverless eventdriven scenario.In addition,to speed up the assessment,a task orchestration centre is designed to split the assessment service into finegrained assessment tasks for parallel execution.Based on the above design scheme,the serverless-based iris recognition measurement system is implemented.Finally,the system is tested for functionality and performance,and various iris recognition algorithms are evaluated using the evaluation system.The results of the tests met the expectations and validated the feasibility of the iris recognition evaluation system.