Fig. 1.

(a) Representation of the 6 first eigenlenses obtained, for each eigenlens k, as $\overline{\mathbf{l}}+a_k{\mathbf{e}}_k$ . Three different $a_k$ values are illustrated for every eigenlens k (in blue, red and yellow). The average lens $\overline{\mathbf{l}}$ is represented in all subfigures in red ( $a_k$ = 0) and the different eigenlenses can be interpreted as deformations from this average lens shape. (b) Example of how the representation of a given lens improves as more eigenlenses are included. The black lens shows the lens ${\mathbf{l}}_n$ to be represented. Blue, red, yellow and green lenses show the obtained representation as the average lens $\overline{\mathbf{l}}$ , the $a_1=-25$ , $a_2=-15$ and $a_5=-5$ coefficients are included, respectively. Asymmetric terms ( $a_3$ and $a_4$ ) are not included in the example for clarity. (c) Representation of the 6 first eigencenters obtained, for each eigencenter k, as $\overline{\mathbf{z}}+c_k{\mathbf{g}}_k$ . Three different $c_k$ values are illustrated for every eigencenter k (in blue, red and yellow). The average central area $\overline{\mathbf{z}}$ is represented in all subfigures in red ( $c_k$ = 0) and the different eigencenters can be interpreted as deformations from this average central area.