摘要
为了提高图像分类的准确度,提出基于最小Hausdorff距离的多示例多标记K近邻图像分类方法。该方法通过改善图像包的生成方法,均匀分割并提取图像的颜色和纹理特征,使用最小Hausdorff距离作为包间的距离度量,对多示例多标记K近邻算法进行改进。实验结果表明,该方法提高了分类准确度,减少了运行时间。
In order to improve the accuracy of image classification,we put forward a method called multi-instance multi-label KNN based on minimal Hausdorff distance. This method uses the minimal Hausdorff distance to measure the distance between bags,and improves the multi-instance multi-label KNN algorithm by improving the generation of image bags,segmenting images on average,and extracting image color and texture features. Experimental results show that the method reduces the running time and improves the classification accuracy.
出处
《济南大学学报(自然科学版)》
CAS
北大核心
2015年第4期275-279,共5页
Journal of University of Jinan(Science and Technology)
基金
国家自然科学基金(61373081)
山东省科技攻关计划(2013GGX10125)
关键词
图像分类
多示例多标记
K近邻
图像分割
特征提取
image classification
multi-Instance multi-Label
K-nearest neighbor
image segmentation
feature extraction