1: \begin{definition}[Orientation and coherency]\label{def:orientation}

2: The matrix $\mat{J}(X)$ defined in Theorem~\ref{pro:wav:based:orientation} (\ref{def:newtensor}) is called the structure tensor of $X$. Let $\lambda_1,\lambda_2$ be its two eigenvalues. The coherency index of $X$ is defined as

3: \[

4: \chi(X)=\frac{|\lambda_2-\lambda_1|}{\lambda_1+\lambda_2}\;.

5: \]

6: An orientation $\vect{\vec n}$ is any unit eigenvector associated to the largest eigenvalue of $\mat{J}(X)$ or equivalently of $\mat{\Sigma}(c_{i,\vect{k}}^{(\mathcal{R})}(X))$.

7: \end{definition}

8: