With the popularity of the Internet,the data on the network increased rapidly,especially the development of multimedia technology.People need to get more and more image information on the Internet and users urgently need to search for the desired target from massive images quickly,especially in the field of electronic commerce.In this case,users need a kind of search technology that can be used to search the target without inputting the text information of the goods,so content-based image retrieval has been paid more and more attention by researchers.In recent years,with the rapid development of mobile Internet and the popularity of mobile smartphones,searching for information on the mobile terminal has been accepted by more and more users,therefore,the image retrieval based on mobile terminal will be applied widely.In this experiment,a content-based image retrieval system was designed.In this system,the smartphone could upload clothing image to the server and the server performed the feature extraction and similarity retrieval.In this way,some of the most similar results in color and style could be acquired and returned to the smartphone.The whole process could be completed in real time.In order to realize this system,an image segmentation algorithm based on color and texture features was proposed firstly,and the target area of the garment image was extracted.On this basis,a method of quadratic clustering was designed and the images in the sample library were clustered based on color.Then,a feature coding algorithm based on multi-scale spatial segmentation was proposed,and a hash algorithm which was specially used for the similarity matching of the encoding feature was proposed,so that the retrieval of the massive image could be realized in real time.Finally,the system was designed.In this process,the related algorithms were realized on the PC,and the information exchange between smartphone and the PC was realized according to the communication technology under the Android platform,then the system development was completed and tested. |