What would you like to understand?

The reconstruction is formulated as a linear combination of the principal component direction vectors, weighted by the stored 2D coordinates. This is equivalent to multiplying the 2D coordinate vector by a reconstruction matrix whose columns are the principal components.

Conditions: A compressed coordinate vector $\mathbf{z} \in \mathbb{R}^2$ is available.; Reconstruction uses principal component directions $\mathbf{v}_1, \mathbf{v}_2$.; The output is an approximation $\hat{\mathbf{x}}$ in $\mathbb{R}^3$.

The reconstruction is formulated as a linear combination of the principal component direction vectors, weighted by the stored 2D coordinates. This is equivalent to multiplying the 2D coordinate vector by a reconstruction matrix whose columns are the principal components.

Conditions: A compressed coordinate vector $\mathbf{z} \in \mathbb{R}^2$ is available.; Reconstruction uses principal component directions $\mathbf{v}_1, \mathbf{v}_2$.; The output is an approximation $\hat{\mathbf{x}}$ in $\mathbb{R}^3$.