This letter discusses a novel low-power digital CMOS architecture for speaker identification (SI) by combining $k$ -means clustering with Gaussian mixture model (GMM) scoring. We show that $k$ -means clustering at the front-end reduces the dimensionality of speech features to minimize downstream processing without affecting SI accuracy. Implementation of cluster generator is discussed with novel distance computing and online centroid update datapaths to minimize overhead of the clustering layer (CL). The integrated design achieves $6\times $ lower energy than the conventional for SI among ten speakers.

中文翻译：

---

通过集成聚类和高斯混合模型评分进行低功率说话人识别

这封信讨论了一种用于说话人识别 (SI) 的新型低功耗数字 CMOS 架构 $千$ - 表示使用高斯混合模型 (GMM) 评分进行聚类。我们证明 $千$ -means 前端的聚类降低了语音特征的维数，从而在不影响 SI 准确度的情况下最小化下游处理。通过新颖的距离计算和在线质心更新数据路径讨论了集群生成器的实现，以最小化集群层 (CL) 的开销。一体化设计实现 $6\times $ 十个扬声器中比传统的 SI 更低的能量。