A Dataset for Voice-Based Human Identity Recognition

This dataset is divided into two main sub-datasets: samePhrase and differentPhrase. Each speaker has the same label in both sub-datasets. In the samePhrase sub-dataset, a speaker repeats the sentence “Machine Learning 1, 2, 3, 4, 5, 6, 7, 8, 9, 10” ten times. The length of each sample is between seven and ten seconds. For the differentPhrase sub-dataset, each speaker contributed with a phrase selected randomly from different resources such as books, songs lyrics, orone-line texts. Each speaker contributed with ten different samples, the length of each sample inthe differentPhrase sub-dataset does not exceed ten seconds本数据集分为两个主要子数据集：samePhrase（同短语）与differentPhrase（异短语）。所有说话者在两个子数据集中均拥有统一标签。在samePhrase（同短语）子数据集中，每位说话者会将句子"机器学习1、2、3、4、5、6、7、8、9、10"重复录制十次，单条样本的时长介于7至10秒之间。在differentPhrase（异短语）子数据集中，每位说话者的语音素材取自随机选取的短语，来源涵盖书籍、歌曲歌词或单行文本等不同渠道。每位说话者提供10条不同的语音样本，且异短语子数据集中的单条样本时长均不超过10秒。