基于局部特征的商标图像检索方法比较
首发时间:2013-09-04
摘要:基于局部特征的商标图像检索是基于内容的图像检索领域的一个热点,它以其较强的鲁棒性和若干不变性在图像检索领域发挥着越来越重要的作用。本文主要对SIFT(Scale-invariant feature transform)算法和基于hessian矩阵的SIFT改进算法SURF算子(Speeded Up Robust Features)在商标检索中的应用进行了分析比较。论文首先指出了基于局部特征的图像检索的方法,然后分别详细介绍了利用SIFT特征进行商标检索的方法和利用SURF算子进行商标检索的方法,并设计了多组实验来对SIFT和SURF算法分别在性能和效率方面进行了综合性地分析比较,最后给出了两种算法的性能总结。
For information in English, please click here
Comparison of Methods for Trademark Image Retrieval Based on Local Feature
Abstract:Trademark image retrieval based on local feature is a hot field in content-based image retrieval and it plays an increasingly important role in the field of image retrieval because of its strong stability and invariance.This paper focuses on the analysis and comparison about SIFT(Scale-invariant feature transform) algorithm and SURF(Speeded Up Robust Features) algorithm ,improved SIFT algorithm based on hessian matrix. Firstly, the paper presents the method of an image retrieval based on local feature.And then it introduces the two methods of SIFT and SURF specifically and designs multiple sets of experiments to analyze and compare the methods of SIFT and SURF algorithm in terms of performance and efficiency. Finally,the paper points out the final performance of the two algorithms.
Keywords: local image features trademark image retrieval SIFT SURF
基金:
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于局部特征的商标图像检索方法比较
评论
全部评论0/1000