One-Class Fake Speech Detection Based on Improved Support Vector Data Description

Figure 2

The architecture of proposed methods. (a) The architecture of AESVDD-1. (b) The architecture of AESVDD-2.

(a)

(b)