Semi-supervised Learning for Automatic Modulation Recognition Using Haar Time–Frequency Mask and Positional–Spatial Attention

<div>Augmented examples generated by Fmix and HTF mask. Every two columns is a group, there are three groups of augmented examples generated by Fmix and HTF mask, respectively. For each group, the first row is the original two samples. The second row is the binary mask. From left to right is the Fmix mask and HTF mask, respectively. And the third row is the augmented sample generated by Fmix and HTF mask from left to right.</div>

Wireless Communications and Mobile Computing

fig2

Figure 2

Figure 2: Semi-supervised Learning for Automatic Modulation Recognition Using Haar Time–Frequency Mask and Positional–Spatial Attention 