基于HOG-LBP特征的中药饮片图像识别
|
摘要
目的通过对方向梯度直方图(HOG)、局部二值模式(LBP)等特征提取及融合的方法研究,实现有效的中药饮片图像识别。方法分析HOG和LBP特征,进行特征融合,采取支持向量机(SVM)分类算法,在采集整理的中药饮片图像数据集基础上对算法进行训练、测试和改进,从而获得有效的中药饮片多分类模型,并将模型与其他算法进行比较,评价算法的有效性。结果通过对30种中药饮片2927张图像的训练和测试,等价模式下的HOG-LBP融合特征算法的饮片图像识别率达91.16%,优于传统算法。结论等价模式下HOG-LBP融合特征结合SVM分类器的方法具有较高的识别率,可有效应用于中药饮片的识别和分类。进一步提高数据种类和数据量,有助于提高算法的适用性和识别效果。
Objective To achieve effective image recognition of TCM decoction pieces through research on feature extraction and fusion of HOG and LBP. Methods The HOG and LBP features were analyzed, and feature fusion was performed. A support vector machine(SVM) classification algorithm was adopted. The algorithm was trained, tested and improved on the basis of collecting the image data of TCM decoction pieces, so as to obtain an effective multi-classification model of TCM decoction pieces. The model was compared with other algorithms to evaluate the effectiveness of the algorithm. Results Through the training and testing of 2927 images of 30 TCM decoction pieces, the HOG-LBP fusion feature algorithm in the equivalent mode had a segmentation rate of 91.16%, which was superior to the traditional algorithm. Conclusion The HOG-LBP fusion feature combined with the SVM classifier in the equivalent mode has a high recognition rate and can be effectively applied to the identification and classification of TCM decoction pieces. Further improving the type of data and the amount of data can help to improve the applicability and recognition of the algorithm.
Objective To achieve effective image recognition of TCM decoction pieces through research on feature extraction and fusion of HOG and LBP. Methods The HOG and LBP features were analyzed, and feature fusion was performed. A support vector machine(SVM) classification algorithm was adopted. The algorithm was trained, tested and improved on the basis of collecting the image data of TCM decoction pieces, so as to obtain an effective multi-classification model of TCM decoction pieces. The model was compared with other algorithms to evaluate the effectiveness of the algorithm. Results Through the training and testing of 2927 images of 30 TCM decoction pieces, the HOG-LBP fusion feature algorithm in the equivalent mode had a segmentation rate of 91.16%, which was superior to the traditional algorithm. Conclusion The HOG-LBP fusion feature combined with the SVM classifier in the equivalent mode has a high recognition rate and can be effectively applied to the identification and classification of TCM decoction pieces. Further improving the type of data and the amount of data can help to improve the applicability and recognition of the algorithm.
引文
[1]YONG R,HUANG T S,CHANG S F.Image retrieval:current techniques,promising directions,and open issues[J].Journal of Visual Communication and Image Representation,2015,10(1):39-62.
[2]木拉提·哈米提,员伟康,严传波,等.基于颜色直方图的新疆维吾尔医药材图像特征提取与判别分析[J].生物医学工程学杂志,2015,32(3):588-593.
[3]李震.中药饮片特征提取和识别系统[D].哈尔滨:哈尔滨工业大学,2013.
[4]黄道,何晋,刘志愿,等.组合分类器算法bagging和adaboostM1在中草药分类的研究[J].中国新通信,2014,16(4):21.
[5]OJALA T,PIETIKAINEN M,MAENPAA T.Gray scale and rotation invariant texture classification with local binary patterns[J].European Conference on Computer Vision,2000,24(7):404-420.
[6]DALAL N,TRIGGS B.Histograms of oriented gradients for human detection[J].Computer Vision and Pattern Recognition,2005,1(12):886-893.
[7]胡仕玲,顾爽,陈启军.基于HOG的物体分类方法[J].华中科技大学学报:自然科学版,2011,39(S2):124-126.
[8]MARCOS J,FERNáNDEZ R.Computer vision for pedestrian detection using Histograms of Oriented Gradient[D].Boston:Massachusetts Institute of Technology,2014.
[9]FLACH P A.On the state of the art in machine learning:A personal review[J].Artificial Intelligence,2001,131(1):199-222.
[10]陈梦竹,岑翼刚,许家佗,等.基于图像处理的望诊面色自动识别研究[J].中国中医药信息杂志,2018,25(12):97-101.
[2]木拉提·哈米提,员伟康,严传波,等.基于颜色直方图的新疆维吾尔医药材图像特征提取与判别分析[J].生物医学工程学杂志,2015,32(3):588-593.
[3]李震.中药饮片特征提取和识别系统[D].哈尔滨:哈尔滨工业大学,2013.
[4]黄道,何晋,刘志愿,等.组合分类器算法bagging和adaboostM1在中草药分类的研究[J].中国新通信,2014,16(4):21.
[5]OJALA T,PIETIKAINEN M,MAENPAA T.Gray scale and rotation invariant texture classification with local binary patterns[J].European Conference on Computer Vision,2000,24(7):404-420.
[6]DALAL N,TRIGGS B.Histograms of oriented gradients for human detection[J].Computer Vision and Pattern Recognition,2005,1(12):886-893.
[7]胡仕玲,顾爽,陈启军.基于HOG的物体分类方法[J].华中科技大学学报:自然科学版,2011,39(S2):124-126.
[8]MARCOS J,FERNáNDEZ R.Computer vision for pedestrian detection using Histograms of Oriented Gradient[D].Boston:Massachusetts Institute of Technology,2014.
[9]FLACH P A.On the state of the art in machine learning:A personal review[J].Artificial Intelligence,2001,131(1):199-222.
[10]陈梦竹,岑翼刚,许家佗,等.基于图像处理的望诊面色自动识别研究[J].中国中医药信息杂志,2018,25(12):97-101.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |