How to generate APC_feature_base.npy of each person?

[vicdxxx]

dear fellow I found manifold projection use APC_feature_base.npy each person, but not clear how to generate this file. Is that use target person voice to train audio2feature_model?

[vicdxxx]

I found the answer is yes

[YuanxunLu]

1. APC_feature_base.npy is the generated APC features from training audio sequences of each person, manifold projection learns a mapping from input audio to the training database. Please check the paper.
2. audio2feature model is person-specific, you should train each audio2feature model for each person.

[vicdxxx]

这是来自QQ邮箱的假期自动回复邮件。   您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。