摘要
针对鞋印图像结构形态特点,提出一种基于融合特征和极限学习机(ELM)的鞋印图像识别方法,将图像均匀分成152个大小相等的子区域,并提取相应的纹理、形状特征,结合两种特征信息提取子区域融合特征向量,再通过逐行扫描子区域的方法计算出鞋印图像融合特征向量,然后利用ELM实现对鞋印图像的识别。实验结果表明,该方法可以有效对鞋印图像进行识别,比其他特征识别方法具有更高的准确率。
Aiming at the structural and morphological characteristics of shoeprint image, a recognition method of shoeprint image based on fusion feature and extreme learning machine(ELM) is proposed. The image is evenly divided into 152 sub-regions of the same size, and the corresponding texture and shape features are extracted. The fusion feature vector of the sub-regions is extracted by combining the two feature information. Then the fusion feature vector of the shoe print image is obtained through progressive scanning of sub-regions. Then the recognition of the shoeprint image is realized by ELM. The experimental results show that the method used to recognize the shoeprint image is effective, and has higher accuracy than other methods.
作者
杜明坤
王茜仪
蔡星宇
Du Mingkun;Wang Qianyi;Cai Xingyu(Department of criminal science and technology,Jiangsu Police Institute,Jiangsu,Nanjing,210000)
出处
《电子技术(上海)》
2018年第10期7-10,6,共5页
Electronic Technology
基金
江苏警官学院校级科研项目(2016SJYZQ08
2016SJYZQ03)
关键词
融合特征
极限学习机
鞋印图像识别
分区域
fusion feature
extreme learning machine(ELM)
shoeprint image recognition
sub-regions