摘要
针对Trace变换提取的图像特征缺乏对纹理边缘信息描述和计算代价高的问题,利用小波变换对图像轮廓的表征优势,提出了多分辨率Trace变换并应用于纹理图像分类.首先,将小波变换引入到Trace变换中,对纹理图像进行非下采样小波变换,得到不同频率的低频特征子图及高频边缘子图;其次,在各级子图上进行一组泛函的Trace变换,获取纹理图像的融合特征,在获得图像边缘信息的同时避免了Trace变换不同泛函组合计算代价过高的问题;最后,把融合特征送入支持向量机对图像进行分类.实验结果表明,对图像采用多分辨率Trace变换提取的融合特征具有更好的纹理描述能力,相对于传统Trace变换及MCM等对比方法具有更高的鉴别性能,且在时间效率上相对于传统Trace变换有大幅提升.
There is a problem that the image features extracted by trace transform lack description of texture edge information,and the computation cost is high,too.Based on the advantages of wavelet transform in image contour representation,a new fusion feature extraction algorithm,multi-resolution trace transform,is proposed and applied to texture image classification.Firstly,the wavelet transform was introduced in trace transform,low frequency feature sub images and high frequency edge sub images of texture images at different frequencies are obtained by using nonsubsampled wavelet transform.Then,we carried out a set of functional trace transform on each level sub images to obtain the fusion features of texture image,which not only obtains the edge information of the image,but also avoids the problem of high cost.Finally,the fusion features were fed into support vector machines to classify the images.The experiment results show that the fusion features of multi-resolution trace transformation have better texture description ability and achieves higher recognition rate than the original trace transform and MCM contrast method,the time efficiency is greatly improved compared to the traditional trace transform.
引文
[1] 刘丽,匡纲要.图像纹理特征提取方法综述[J].中国图象图形学报,2009,14(4):622-635.LIU Li,KUANG Gang-yao.Overview of image texture feature extraction methods[J].Journal of Image and Graphics,2009,14(4):622-635.(in Chinese)
[2] 迟健男,张闯,张朝晖,等.基于反对称双正交小波重构的图像增强方法[J].自动化学报,2010,36(4):475-487.CHI Jian-nan,ZHANG Chuang,ZHANG Chao-hui,et al.Image enhancement based on anti-symmetrical biorthogonal wavelet recontruction[J].Acta Automatica Sinica,2010,36(4):475-487.(in Chinese)
[3] 周先春,吴婷,石兰芳,等.一种基于曲率变分正则化的小波变换图像去噪方法[J].电子学报,2018,46(3):621-628.ZHOU Xian-chun,WU Ting,SHI Lan-fang,et al.A kind of wavelet transform image denoising method based on curvature variation regularization[J].Acta Electronica Sinica,2018,46(3):621-628.(in Chinese)
[4] 李亚峰.基于图像分解的稀疏正则化多区域图像分割方法[J].电子学报,2015,43(9):1841-1849.LI Ya-feng.A sparsity Regularized multiregion image segmentation method based on image decomposition[J].Acta Electronica Sinica,2015,43(9):1841-1849.(in Chinese)
[5] 李哲涛,李仁发,谢井雄.基于全向小波的图像边缘检测算法[J].电子学报,2012,40(12):2451-2455.LI Zhe-tao,LI Ren-fa,XIE Jing-xiong.An edge detection algorithm based on omni-directional wavelet transform[J].Acta Electronica Sinica,2012,40(12):2451-2455.(in Chinese)
[6] 殷明,白瑞峰,邢燕,等.基于非下采样双树复小波域的双变量模型去噪算法[J].光子学报,2014,43(10):131-137.YIN Ming,BAI Rui-feng,XING Yan,et al.Denoising algorithm by nonsubsampled dual-tree comples wavelet domain bivariate model[J].Acta Photonica Sinica,2014,43(10):131-137.(in Chinese)
[7] KADYROV A,PETROU M.The trace transform and its applications[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2001,23(8):811-828.
[8] KADYROV A,TALEPBOUR A,PETROU M.Texture classification with thousands of features[A].British Machine Vision Conference[C].UK:BMVC,2014.665.
[9] SRISUK S,PETROU M,KURUTACH W,et al.A face authentication system using the trace transform[J].Pattern Analysis and Applications,2005,8(1):50-61.
[10] NASRUDIN M F,PETROU M.Offline handwritten Jawi recognition using the trace transform[A].International Conference on Pattern Analysis and Intelligent Robotics[C].US:IEEE,2011.87-91.
[11] GOUDELIS G,KARPOUZIS K,KOLLIAS S.Exploring trace transform for robust human action recognition[J].Pattern Recognition,2013,46(12):3238-3248.
[12] ALBUKHANAJER W,BRIFFA J,JIN Y.Evolutionary multi-objective image feature extraction in the presence of noise[J].IEEE Transactions on Cybernetics,2014,45(9):1757-1768.
[13] PETROU A M,WANG F.A tutorial on the practical implementation of the trace transform[A].Handbook of Texture Analysis[M].UK:Imperial College Press,2008.313-346.
[14] 秦定宇,王敬东,李鹏.图像融合中小波基的选择分析[J].光电子技术,2006,26(3):203-207.QIN Ding-yu,WANG Jing-dong,LI Peng.Wavelet base selection and evaluation for image fusion[J].Opto Electronic Technology,2006,26(3):203-207.(in Chinese)
[15] 钟桦,杨晓鸣,焦李成.基于多分辨共生矩阵的纹理图像分类[J].计算机研究与发展,2011,48(11):1991-1999.ZHONG Hua,YANG Xiao-ming,JIAO Li-cheng.Texture classification based on multiresolution co-occurrence matrix[J].Journal of Computer Research and Development,2011,48(11):1991-1999.(in Chinese)
[16] 徐卓飞,张海燕,刘凯,等.基于Radon-经验模式分析的纹理分类[J].中国图象图形学报,2015,20(8):1091-1101.XU Zhuo-fei,ZHANG Hai-yan,LIU Kai,et al.Texture classification based on Radon-empirical mode decomposition analysis[J].Journal of Image and Graphics,2015,20(8):1091-1101.(in Chinese)
[17] CROSIER M,GRIFFIN L D.Using basic image features for texture classification[J].International Journal of Computer Vision,2010,88(3):447-460.
[18] 周林,平西建,童莉.基于改进基本图像特征直方图的纹理分类算法[J].系统工程与电子技术,2012,34(6):1272-1277.ZHOU Lin,PING Xi-jian,TONG Li.Texture classification based on improved basic image feature histogram[J].Systems Engineering and Electronics,2012,34(6):1272-1277.(in Chinese)
[19] 刘金平,唐朝晖,徐鹏飞,等.基于B样条概率密度估计的纹理图像分类[J].光电子·激光,2017,(5):538-546.LIU Jin-ping,TANG Zhao-hui,XU Peng-fei,et al.On texture image classification via B-spline probability density function estimation[J].Journal of Optoelectronics·Laser,2017,(5):538-546.(in Chinese)
[20] 郑建炜,杨平,王万良,等.组加权约束的核稀疏表示分类算法[J].计算机研究与发展,2016,53(11):2567-2582.ZHENG Jian-wei,YANG Ping,WANG Wan-liang,et al.Kernel sparse representation classification with group weighted constraints[J].Journal of Computer Research Development,2016,53(11):2567-2582.(in Chinese)