摘要
提出了一种利用Gabor滤波器组对图像进行纹理特征提取,然后将这些特征向量作为外部刺激输入给PCNN对图像进行分割的新方法。该方法既保持了精确的分割结果,同时又解决了Gabor滤波器运算数据量大处理速度慢的问题。从实验结果可以看出,该方法与传统采用Gabor的纹理分割方法相比速度有很大的提高,而其分割精度与传统分割方法相似。
An algorithm based on Gabor filters is presented to extract image texture character. The extracted features are inputted into the PCNN to segment the image. This method can achieve good segmentation result and improve the algorithm＇s processing speed. The experiment results show that the proposed method process faster than the traditional method using Gabor filters, meanwhile, both of them have the similar segmentation accuracy.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
2010年第1期11-15,共5页
Journal of University of Electronic Science and Technology of China
基金
国家自然科学基金(60573019)
广东省自然科学基金(07300561)
关键词
特征提取
滤波器
脉冲耦合神经网络
纹理分割
feature extraction
filters
pulse coupled neural network
texture segmentation
