Design And Implementation Of Automotive Information Labeling System Based On SSM Framework

Due to the advantages of product functions and performance brought by machine learning technology,the application prospects are increasingly broad in different industry backgrounds.In machine learning engineering,the quality of the data used for training greatly affects the effectiveness of the model being trained.In order to facilitate the algorithm model to obtain training data with obvious features,the original data needs to be annotated to facilitate the algorithm model to discover the patterns and rules therein.In this context,the automotive information labeling system is to process the car information related image,voice and text data,and convert the car information raw data into a data set for algorithm model training through full manual or semi-automatic labeling.According to the usage scenario and business needs,the system selects the SSM(Spring + Spring MVC + My Batis)framework as the main development framework.From a functional perspective,the system is divided into four modules: user management,data processing,task management,and user training.According to the characteristics of huge amount of data in the automobile segmentation field and diversified evaluation indicators,the core functions of product labeling,collection and review are implemented on the webpage and mobile terminal through the data processing module.At the same time,in order to facilitate the management of tagging,collection and review personnel and system tasks,and reduce the workload of user training,user addition modules can be implemented through the user management module,permission management and performance query,the Kanban module monitors task progress,and the task training module enables user training.In order to support the semi-automatic labeling data transmission function,according to the permission control requirements of the intranet interface,Shiro security control framework is used to implement interface identity permission control,and complete the interaction between system data and algorithm models.In order to ensure system data security,according to the characteristics of a single user name corresponding to multiple user entities in the actual environment,for high-secure tasks,the self-signed certificate generated by the Keytool tool is stored in the hardware device,and the public key cryptosystem is used to implement the two-way authentication method to achieve system operation The two-way identity authentication between the server and the server,and the data transmission process is realized through symmetric encryption.The automotive information labeling system obtains the required data sets for machine learning algorithm models through data labeling,manual review,and the use of data collection tools.Through testing,the basic functions of the car information marking system have been fully realized,and the system can run normally and meet the expected results.