Algorithm 1 The multi-level quantization algorithm based on PPF

Input: The normalized feature vector ${\overline{H}}_{\alpha ,k}^{\prime }$; the quantization factor $\delta$ with a default value: 10${}^{-2}$ Output: The quantized bit sequence $Q_{\alpha }$;   1:Calculate the mean $\mu$ and the standard deviation $\sigma$ of the feature vector ${\overline{H}}_{\alpha ,k}^{\prime }$;   2:Use PPF of the $scipy.stats$ library ($ppf$) to confirm the benchmarks;   3:$lo{w}_1=ppf(0.25-\delta ,loc=\mu ,scale=\sigma )$;   4:$hig{h}_1=ppf(0.25+\delta ,loc=\mu ,scale=\sigma )$;   5:$lo{w}_2=ppf(0.75-\delta ,loc=\mu ,scale=\sigma )$;   6:$hig{h}_2=ppf(0.75+\delta ,loc=\mu ,scale=\sigma )$;   7:Initialize $Q_{\alpha }$ as an empty list;   8:for$iinrange\left(len\left({\overline{H}}_{\alpha ,k}\right)\right)$do   9:  if ${\overline{H}}_{\alpha ,k,i}^{\prime }>=lo{w}_1$ and ${\overline{H}}_{\alpha ,k,i}^{\prime }<=hig{h}_1$ then 10:   $Q_{\alpha }.append\left(0\right)$ 11:  else 12:   if ${\overline{H}}_{\alpha ,k,i}^{\prime }>=lo{w}_2$ and ${\overline{H}}_{\alpha ,k,i}^{\prime }<=hig{h}_2$ then 13:    $Q_{\alpha }.append\left(1\right)$ 14:   else 15:    $Q_{\alpha }.append(-1)$ 16:   end if 17:  end if 18:end for 19:return $Q_{\alpha }$.