Standard PC-GITA (s-PC-GITA)

该数据集包含了100位个体的录音，其中两组各占一半：一组是50位帕金森病（PD）患者，另一组是50位健康对照者（HC）。每组中男女各占25人。录音是在隔音室内进行的，包括音节快速重复（DDK）练习、朗读句子和独白。为了本研究，原始录音从44.1千赫兹采样率降低到了16千赫兹。该数据集的规模为100人（50名PD患者和50名HC对照者），任务是：从语音中检测帕金森病。

This dataset contains speech recordings from 100 participants, split into two equal groups: 50 patients with Parkinson’s disease (PD) and 50 healthy controls (HC). Each group has an equal gender distribution, with 25 males and 25 females respectively. All recordings were collected in a soundproof room, including three types of speech tasks: rapid syllable repetition (DDK) exercises, sentence reading, and monologue. For this study, the original recordings were downsampled from a 44.1 kHz sampling rate to 16 kHz. The core task associated with this dataset is to detect Parkinson’s disease from speech signals.

搜集汇总

背景与挑战

背景概述

该数据集研究帕金森病语音检测中可解释性方法的有效性，包含多种语音任务的样本解释，结果表明现有方法在提供临床价值方面存在局限性。