The Clothing Image Retrieval System Design Based On Mobile Vision

With the development of mobile Internet era, the intelligent mobile phoneshopping come into fashion. For the mobile phone online apparel shopping systemonly allow the user to use the brand, price, color text information to search. Thispaper attempts to use the real-time, rich image information of the mobile terminalcharacteristics to search clothing. Therefore this paper proposes image of clothingSURF image registration algorithm based on search and matching system. Thesystem cooperate with magazines, newspapers, and other traditional media, using itsclothing image source as the experiment with the source image database. And thenthrough the SURF feature extraction algorithm of clothing image store in thedatabase, and store the clothing with corresponding information. The mobile phoneclient takes clothing photo and upload, then server returns search results costumeretrieval based on SURF algorithm.The SURF algorithm has scale variation, rotation, affine transform, noiserobustness, at the same time in the real-time retrieval performance. Therefore, weuse the SURF algorithm to complete the characteristics of clothing image detection,feature description, feature matching work, according to the SURF algorithm, theproblem may exist point error, error matching filter and elimination of feature pointsusing the RANSAC algorithm. At the same time for the poor performance problemsof the data in the high concurrency IO, we put forward garment feature data into thecache solution. Finally, for the latency problem of the image of clothing in themobile phone client show, we put forward to show thumbnail generation scheme.In the simulation of the Internet environment for the relevant retrievalexperiments, experiments show that the system can accurate efficient for clothingimage retrieval and matching, and has high matching rate. Mobile visual clothing retrieval system enhances the convenience and fast retrieval system based on theclothing, and it is the supplement and exploration in the context of mobile internet.