BasCom
收藏资源简介:
BasCom数据集是由比勒费尔德大学技术学院创建的新型表面肌电信号数据集,专门用于零样本学习研究。该数据集包含11名参与者采集的11种基本动作和8种组合动作,规模超越了先前同类数据集,为肌电信号分析提供了更丰富的样本基础。数据采集过程通过标准表面肌电信号记录设备完成,经过严格的时间窗口分割和多通道信号预处理,形成了结构化的时间序列数据。该数据集主要应用于智能假肢控制领域,旨在解决组合动作识别中的校准负担问题,为零样本学习算法在临床康复应用中的性能评估提供关键实验数据。
The BasCom dataset is a novel surface electromyography (sEMG) dataset developed by the Faculty of Technology, Bielefeld University, exclusively for zero-shot learning research. It contains 11 basic movements and 8 combined movements collected from 11 participants, with a scale surpassing that of prior comparable datasets, thus providing a more abundant sample foundation for sEMG signal analysis. The data collection was conducted using standard sEMG recording devices, and the raw signals were subjected to rigorous time window segmentation and multi-channel signal preprocessing to generate structured time-series data. This dataset is primarily utilized in the field of intelligent prosthetic control, aiming to address the calibration burden in combined movement recognition and provide key experimental data for evaluating the performance of zero-shot learning algorithms in clinical rehabilitation applications.
数据集概述
该数据集详情页面是一个GitLab群组,包含与“sEMG零样本学习”相关的多个子项目,具体如下:
- BasComdataset:一个项目,已创建8个月。
- sEMGsampling:一个项目,已创建8个月。
- Zeroshotmethods:一个项目,已创建8个月。
该群组主要用于组织和托管与“sEMG”(表面肌电图)和“零样本学习”(Zero-shot learning)相关的代码或数据集资源。

- 1Prototype Adaptation for Zero-Shot sEMG Movement Classification比勒费尔德大学·技术学院 · 2026年



