Design And Implementation Of Content-based Image Recognition And Search System

With the rapid development of the Internet,picture reading has become an important way for people to consume information on the Internet.The search engine's understanding of picture needs to be transformed from single text to image content,help users master the internal information and provide relevant images which becomes a demand.However,all kinds of online image retrieval systems cannot meet the diversified needs of users.Therefore,it is very important to develop a full-featured image content search system to provide accurate and efficient image recognition services.The Image Recognition and Search System is based on the company's image search department's massive Internet image information data and advanced image recognition technology of Institute of Deep Learning(IDL).Based on the detailed study of the image recognition search system at home and abroad,proposes a complete map of the local and Internet images,and then user uploads the images and extracts the image features of the images.At the same time,the classification picture information database retrieves the whole network related picture information.Secondly,analyzes the requirements of the system and divides the functional modules such as keyword extraction,multi-size picture display,picture source acquisition,similar picture recommendation,similar face encyclopedia,multi-semantic label expansion and similar commodity retrieval.Based on the internal ODP framework which make the realization of multi-module information data integration,based on React to provide a modular and one-way data flow front-end integrated system and finally displayed to the user.At the end,the functions of the system are fully realized and the indexes of function and performance are rigorously tested.At the same time,the function,effect and performance iterations are optimized by combining the active and passive feedback information of the users,which improves the accuracy of image recognition and the fast response processing.ability.Compared with the previous application,the system has high image recognition accuracy,more complete image retrieval and faster server processing efficiency,which can break the technical barrier of foreign knowledge map and improve the user experience.The design and implementation of the system are reasonable and feasible,and have good effect,which can be used for reference to other system.