Download presentation
Presentation is loading. Please wait.
1
基于人眼追踪的手机解锁系统 报告人:李映辉 指导老师:王继良
Unlock with A Glance 基于人眼追踪的手机解锁系统 报告人:李映辉 指导老师:王继良
2
Background Mobile devices hold sensitive information
Authentication schemes are prone to shoulder-surfing attacks Unlock with biological information sometimes fails
3
Related Works GazeTouchPass EyeDent On mobile devices
Use gaze direction and touch to unlock phone Only detect gaze directions EyeDent Based on computer Use dedicated eye tracker
4
Related Works Gaze Tracking Method Accuracy(°) Speed(fps) Limitation
CHM Mapping (1.9, 2.2) NA Based on computers EyeTab Model Based 6.88±1.8 12 Low Accuracy Pupil-Canthi-Ratio Gaussian Process Regression 3.9 Not accurate in certain area GT on Tablet ANN 4.42 0.7 Too slow
5
Design Pupil Detection Gaze Estimation Mapping Authentication
6
Challenges & Potential Solutions
Gaze estimation is not accurate enough Use Cluster Partition method to filter wrong data Speed is important Fine-grained pupil detection is the bottleneck Use mapping-based gaze tracking technique
7
Timeline 周次 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 工作安排 文献阅读 方案设计
眼球追踪算法实现 解锁功能实现 中期答辩 后期性能优化 撰写论文 期末答辩
8
THANKS FOR LISTENNING Q & A
Similar presentations