利索-全球专利检索

  1. 登录
  2. 注册
发明公开
CN107704822A 基于残缺补全的指静脉深层特征提取与匹配方法及系统 无效 - 撤回

基于残缺补全的指静脉深层特征提取与匹配方法及系统
摘要：
本发明公开了基于残缺补全的指静脉深层特征提取与匹配方法及系统，其特征在于，指静脉深层特征提取与匹配方法包括：对待注册指静脉图像进行预处理，获取预处理后的待注册指静脉图像；采用深度卷积神经网络对预处理后的待注册指静脉图像进行深层特征提取训练处理，获取待注册指静脉图像深层特征；采用待注册指静脉图像深层特征进行注册并构建模板库处理，获取注册后的留存深层特征模板库；采集待识别指静脉图像，将待识别指静脉图像进行特征提取，获取待识别指静脉图像深层特征；采用待识别指静脉图像深层特征与注册后的留存深层特征模板库进行匹配识别处理，获取匹配识别结果。在本发明实施例中，提取出指静脉图像的高区分度特征，提高匹配精度。
摘要（英）：

The invention discloses a finger vein deep feature extraction and matching method and system based on incomplete complement. The finger vein deep feature extraction and matching method is characterized by comprising: preprocessing a finger vein image to be registered to obtain a preprocessed finger vein image to be registered; performing deep feature extraction training on the preprocessed fingervein image to be registered by adopting a deep convolutional neural network to obtain deep features of the finger vein image to be registered; performing registration by adopting the deep features ofthe finger vein image to be registered and constructing a template library for processing to obtain a registered template library reserving the deep features; acquiring a finger vein image to be recognized, extracting features from the finger vein image to be recognized to obtain deep features of the finger vein image to be recognized; and matching the deep features of the finger vein image to berecognized with the registered template library reserving the deep features to obtain a matching recognition result. By extracting high distinguishing features of the finger vein image, the matching precision is improved.

0/0