计算机应用与软件Issue(10)：102-104,107,4.DOI:10.3969/j.issn.1000-386x.2013.10.028
基于改进的LBP方法相结合的尿液细胞识别研究
ON URINE CELLS RECOGNITION BASED ON IMPROVED LOCAL BINARY PATTERN
摘要
Abstract
The effects of urine cells identification and classification using support vector machine (SVM)in two colour coordinate systems of RGB and HSI respectively are analysed and compared in this paper.The effects of comprehensive recognition of the classified urine cells using colour feature parameter and texture feature parameter are also analysed and compared.We also propose a texture feature extraction method which is based on improved local binary pattern (LBP).Experimental results show that the method combining the HSI colour feature, the improved LBP-based texture feature and the SVM has good effect in urine cells recognition and classification.关键词
支持向量机/图像处理/机器学习/局部二值模式(LBP)/目标识别Key words
Support vector machine (SVM)/Image processing/Machine learning/LBP/Target recognition分类
信息技术与安全科学引用本文复制引用
秦颖博,孙杰..基于改进的LBP方法相结合的尿液细胞识别研究[J].计算机应用与软件,2013,(10):102-104,107,4.基金项目
天津市科委自然科学基金项目(06 YF ,JMC15600). （06 YF ,JMC15600）
