Font Size: a A A

Design And Implementation Of Content Based Apparel Retrieval System

Posted on:2019-11-05Degree:MasterType:Thesis
Country:ChinaCandidate:N N LiuFull Text:PDF
GTID:2428330593451694Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
Apparel image retrieval technology belongs to the emerging field of traditional clothing retrieval and e-commerce.With the rapid development of internet and users' purchasing power in e-commerce websites,apparel retrieval technology has also become a research hotspot.How to use efficient and efficient ways to help users find the favorite clothing is currently a very important task.The traditional text-based clothing retrieval method has obvious drawbacks.First of all,all image data in the database based on the text retrieval method need to be annotated manually,and there is a huge workload.Secondly,due to the subjectivity of the human,the label of the artificial annotation is inaccurate and the accuracy of the retrieval result is not high.On the other hand,for the users,sometimes they can not accurately describe the names of the related garments and can not accurately search;these are the drawbacks of the traditional garments retrieval.The content-based clothing retrieval technology is different from the previous text-based retrieval methods,which use the underlying features of the image retrieval,greatly improving the retrieval accuracy.The system is divided into three parts: the client,server and database.The client is responsible for receiving the user upload target image,the server is responsible for processing the retrieval and return results,the database is responsible for storing garment image data.The system uses the current Spring framework to build server MVC system architecture,using the most popular database Mysql storage clothing image data.Firstly,the color feature,the LBP texture feature and the SIFT feature are extracted from the clothing image.At the same time,the image of the clothing image uploaded by the user is also extracted.Similar clothes are retrieved through the similarity calculation and returned to the client.In this thesis,the apparel retrieval system is deeply analyzed based on the software engineering design ideas.The main web framework is used to design and implement a clothing retrieval system with a total code of 18000.At present the main function of the system and the test is completed,and the feasibility of the system is verified through experiments.
Keywords/Search Tags:Clothing Image Retrieval, Feature Extraction, Lucene
PDF Full Text Request
Related items