Figure 1.

Plot of the maximum margin linear decision function $D\left(\mathit{x}\right)={\mathit{w}}^{\top }\mathit{x}+\mathit{b}(\mathbf{\phi }\left(\mathit{x}\right)=\mathit{x})$ (left) and an image of it (right). The gray levels encode $\left|D\right(\mathit{x}\left)\right|$, solid black corresponds to $D\left(\mathit{x}\right)=\mathbf{0}$, and the numbers indicate the supporting vectors (Source: [34]).