首页>
外国专利>
SIFT FEATURE BAG BASED BOVINE IRIS IMAGE RECOGNITION METHOD
SIFT FEATURE BAG BASED BOVINE IRIS IMAGE RECOGNITION METHOD
展开▼
机译：基于SIFT特征包的牛眼虹膜图像识别方法
展开▼
页面导航
摘要
著录项
相似文献
摘要
Disclosed are a SIFT feature bag based bovine iris image recognition method, comprising the following steps: preprocessing an iris image to obtain and effective region; obtaining feature points by using a SIFT method; positioning an inner edge by using an active contour line method; removing the feature point in the inner edge to obtain an effective SIFT feature point set; performing comparison with an optimal SIFT feature bag to obtain a feature histogram; calculating the histogram distance between a to-be-recognized image and each image in a target iris library, and using an object corresponding to the target bovine iris image with the smallest histogram distance as an recognition result. The present invention can accurately perform recognition in cases of that the to-be-recognized bovine iris image is rotated, shifted, partially blocked, or inconsistent in scale, so as to improve the accuracy and reliability of bovine iris image recognition, thereby promoting application of the iris-based recognition method in the food traceability system.
展开▼
