Subject-wise classification accuracies by using CNN, LSTM, Bi-LSTM, proposed stack, and fft methods for the classification of a 2-class hand-gripping HBO-fNIRS data.

Subject-wise classification accuracies by using CNN, LSTM, Bi-LSTM, proposed stack, and fft methods for the classification of a 2-class hand-gripping HBO-fNIRS data.针对二分类手部抓握任务的含氧血红蛋白功能近红外光谱（HBO-fNIRS）数据分类任务，采用卷积神经网络（CNN）、长短期记忆网络（LSTM）、双向长短期记忆网络（Bi-LSTM）、所提出的堆叠模型以及快速傅里叶变换（FFT）方法得到的逐受试者分类准确率。