Continuous Human Motion Recognition Based on FMCW Radar and Transformer

Figure 3

The final tensor dimension of the original data through the feature extraction layer.

(a) Image tensor after feature extraction

(b) A single feature sequence