A Novel Time-Incremental End-to-End Shared Neural Network with Attention-Based Feature Fusion for Multiclass Motor Imagery Recognition

Figure 11

Comparison results of the accuracy and kappa coefficient. (a, c) The comparison results between the traditional nondeep learning methods and our method in this paper. (b, d) The comparison results between the deep learning methods and our method in this paper. It is obvious from (a, b) that our method has the minimum individual difference among different subjects, and also as can be seen from (c, d) our method has greatly improved the overall accuracy of all 9 subjects.

(a)

(b)

(c)

(d)