Design And Implementation Of Auxiliary Function Of Sensitive Image Recognition System

In recent years,video communication has become a reality with the continuous development and improvement of streaming media technology.It spawns a large number of network broadcast platform and becomes a new kind of Internet industry due to its wide audience,obtain diversity and highly interactive.This paper designs and implements the auxiliary function of sensitive image recognition system.It makes the network broadcast content meet the nouns based on the full use of existing laboratory sensitive image recognition algorithm.Moreover,it provides a positive network broadcast environment for the public and a regulated environment for the development of the network broadcast the emerging Internet industry.The auxiliary function of sensitive image recognition system adopts a modular design scheme.It improves the scalability of the system and facilitates the access of subsequent algorithm and function.By analyzing the characteristics of the sensitive image recognition algorithm and user characteristics,the author independently designs and implements the following five modules:Application startup module:The main function of the module is to edit the Http message,send the Http request in POST mode,call the image recognition API module concurrently,and output the image recognition result in the form of Log.In the application startup module,the message sent to the image recognition API module can be selected in the two ways of image data and image URL.At the same time,the module can set the number of images that are submitted to the image recognition API module each time according to the parameters.Image recognition API module:This module is a monitoring module that receives Http requests from the application startup module.The module has the following functions:use the key hash message authentication code(Hash-based Message Authentication the Code,referred to as HMAC)to verify user information;use the Apache Thrift framework,through the connection pool module remote call module and image recognition algorithm according to the results of image classification;return the JSON format recognition results to the application startup module.Image preservation module:The image files that are trusted within the scope of the doubtful threshold are saved to the cloud storage platform.Asynchronous update module:The related information about the suspicious image is stored in asynchronous mode so that the staff can make artificial recognition of the suspicious image and use MySQL as the storage database.Connection pool module:Create and manage connection with image recognition algorithm module,so that connection can be efficiently reused,reduce time consumption and resource consumption when connection is created and closed.The system test results show that the functional requirements of the auxiliary function of sensitive image recognition system have been met.At present,this version is under iterative development and testing process temporarily,while the designing and optimizing of the software still in progress.When the project launches,the company will take this system as the basis,to provide a high quality live entertainment environment for the public,to bring on a healthy and positive development space to the company.