Target sample mining with modified activation residual network for speaker verification.

As can be seen from Fig 3 : (1) As the number of epoch increases, the error rate (EER) and minDCF for all three proposed methods show a decreasing trend under different parameters.

also in 47,180 other papers