摘要
局部二值化模式是对局部信息进行特征提取,适合解决纹理特征不明显的图像特征提取问题,但这种方法存在特征维数高的问题。针对此问题,提出一种适用于半色调图像特征提取的邻域相似性描述子方法。该方法将8个邻域分别与中心像素值进行比较运算,然后求这8个运算值间的相似性指标。相似性指标的统计量进行归一化后作为图像的纹理特征向量。试验中,采用BP神经网络对提取图像特征进行分类试验。试验结果表明,在计算复杂度、识别精度等方面,本研究提出的方法优于单纯使用局部二值化模式算法。
Local binary pattern was obtained by extracting local features,which was appropriate for unobvious textural features but suffered high feature dimension problem.To solve this problem,the neighborhood similarity descriptor method for halftone image feature extraction was proposed.First,the center pixel was compared with its eight neighbor-hood pixels.Second,the similarity index was computed for those pixel pairs.The similarity index was taken as textural feature vector after normalizing the statistics.Finally,BP neural network was adopted to classify the extracted image features in experiment.Experimental results showed that the proposed method was better than the local binary pattern al-gorithm in the computational complexity and recognition accuracy.
出处
《山东大学学报（工学版）》
CAS
北大核心
2016年第3期58-64 73,73,共8页
Journal of Shandong University（Engineering Science）
基金
国家自然科学基金资助项目(61170102)
湖南省研究生科研创新资助项目(CX2015B566)
湖南省自然科学基金资助项目(2014JJ2115
015JJ2046)
