pyriemann.utils.distance.distance_logdet¶

pyriemann.utils.distance.distance_logdet(A, B)[source]¶

Log-det distance between two covariance matrices A and B.

$$d = \sqrt{\left(\log(\det(\frac{\mathbf{A}+\mathbf{B}}{2})) - 0.5 \times \log(\det(\mathbf{A}) \det(\mathbf{B}))\right)}$$

Parameters:	A – First covariance matrix B – Second covariance matrix
Returns:	Log-Euclid distance between A and B