Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

some questions #119

Open
JH-Lam opened this issue Aug 18, 2023 · 0 comments
Open

some questions #119

JH-Lam opened this issue Aug 18, 2023 · 0 comments

Comments

@JH-Lam
Copy link

JH-Lam commented Aug 18, 2023

hi all, appreciate your great job on this open project.

经过测试几个场景后,
测试结果:
1)模糊照片容易误判为fake
2)对屏幕拍摄,正确识别为fake
3)对纸张拍摄(没纸张边缘),倾向误判为real
4)对纸张拍摄(有纸张边缘),正确识别为fake
5)Replay videos Review images
A)对于4_0_0_80x80_MiniFASNetV1SE.pth 利用人脸ROI外围区域信息更多,检测fake效果更好;对于real情况二个模型相当于。
B)至少二次摄像以上才会识别假脸。如果第二次是只是简单转发其它人,因为图片本身没变(至少大多数像素)所以肯定还是识别为真人。
C)而且不能脱离手机、拿挂纸张等检测环境,即如果第二次只是对屏幕、纸张等拍摄,然后识别,极有可能还是‘真脸’(还是外围区域起到辅助作用。当把人脸边缘信息不暴露在二次成像时,置信度明显偏低)。其中屏幕欺诈场景识别准确率最低,是不是没有数据进行训练?
D)成像次数越多越容易正确识别。

附注:使用的model是2.7_80x80_MiniFASNetV2.pth, 4_0_0_80x80_MiniFASNetV1SE.pth

总的来说感觉傅立叶技术在此处作用不大?好像直接训练个分类器即可。。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant