You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Pretrained model is available at https://github.com/wy1iu/sphereface#models. Please see the test folder for getting the face embedding. It is the same as the inference of any CNN pretrained model.
In the evaluation script in the test folder, you extract the features of the last fully connected layer as usual but you also extract features of the flipped image and you concatenate these two representations to get the final feature vector.
Could you please explain why ?
I can't get good verification results although I have good identification results.
Is there any pretrained model using this sphereface concept? If so how do i get the face embeddings of a custom image using that pre trained model?
The text was updated successfully, but these errors were encountered: