Trademark Retrieval Method Based on Deep Feature Clustering
The registration of trademarks needs to avoid the registration of identical or relatively similar trademarks.How to quickly and accurately retrieve trademarks with high similarity to the trademark dataset from the massive trademark data has become particularly important.In traditional trademark retrieval methods,the clustering algorithm for trademark features has been improved to average the data in the center points obtained.Trademarks with high similarity are divided into the same center point and each center point corresponds to an average amount of data.Comparing the similarity among similar center points to perform trademark retrieval tasks can achieve good retrieval performance when applied to a large trademark database.
NETL
NSTL
万方数据
