Closed joeyan closed 3 months ago
Previously the 2D covariance matrix was represented with four values:
$$ \Sigma_{2D} = \begin{bmatrix}a & b\\c & d\end{bmatrix} $$
The b and c terms are redundant since the covariance matrix is always symmetric. Instead, we can adopt matrix representation of a conic where:
b
c
$$ A = \begin{bmatrix}a & b/2\\b/2 & c\end{bmatrix} $$
Previously the 2D covariance matrix was represented with four values:
$$ \Sigma_{2D} = \begin{bmatrix}a & b\\c & d\end{bmatrix} $$
The
b
andc
terms are redundant since the covariance matrix is always symmetric. Instead, we can adopt matrix representation of a conic where:$$ A = \begin{bmatrix}a & b/2\\b/2 & c\end{bmatrix} $$