CvgTalk-Lulu-Periocular Recognition
2015-11-12 01:08
253 查看
Iris region is small, esp. from distance
combine periocular region to increase detection region
frontal face
high res. eye region image
issues
only 2D
2D LBP
to obtain 2D periocular region, first register 2D with 3D shape using facial landmark
3D shape matching, ICP
distance matching score between probed periocular 3D point cloud and that from gallery
fusion of 3D and 2D
3D face dataset
105 subject
4,666 face scans
Intra-class variations
combine periocular region to increase detection region
Current status
mostly use ideal faces and iris imagesfrontal face
high res. eye region image
issues
only 2D
Proposed method
3D shape2D LBP
to obtain 2D periocular region, first register 2D with 3D shape using facial landmark
3D shape matching, ICP
distance matching score between probed periocular 3D point cloud and that from gallery
fusion of 3D and 2D
Experiment
Bosphorus DB3D face dataset
105 subject
4,666 face scans
Results
3D > 20% better than 2D textureIntra-class variations
相关文章推荐
- 8,石头,剪刀,布的逻辑代码
- 【reversing.kr】ImagePrc 逆向分析
- 虚拟机 的 网络配置
- PrefabType 预置类型
- 记一次Redis被攻击的事件
- iOS通过dSYM文件分析crash
- OracleApps Dropship 流程
- Q-4
- Android SDK Manager国内无法更新的解决方案
- 新的博客已经启用,欢迎大家访问(402v.com)
- 详解SESSION与COOKIE的区别
- AngularJS实现跨域请求
- 7大因素让男人短寿
- [POJ 1330]Nearest Common Ancestors[LCA](O(dep[u] + dep[v]))
- 解决unbuntu14.04上的eclipse自动退出的问题
- 解决unbuntu14.04上的eclipse自动退出的问题
- android中的启动模式
- Facecat的iOS自学笔记
- js中json字符串转换为对象以及转换是报 缺少";"错误的解决方法
- c语言:编写冒泡排序,排序一个整形数组(从小到大)