Tnorm利用大量冒充者模型来得到均值u和方差δ，再用这两个数据带入到score =(score-u)/δ得到规整化之后的分数。这样使得结果更加合理。-Tnorm impostor model to get mean u and variance δ, then these two data into the score ' = (score-u)/δ score after regularization. This makes the results more reasonable.

对于说话人语音识别得到的分数，先进行2次znorm规整，再做一次tnorm的规整化，使得分数更加合理。-Speaker voice recognition score, the first 2 times znorm structured, do it again regularization tnorm, more reasonable scores.