In this paper is presented a novel image processing tool for the extraction of geometric information in otolith images. It relies on the reconstruction of individual otolith shape histories from otolith images. Based on the proposed non-parametric level-set representation of otolith shape history, applications to the extraction of growth axes and ring structures in otolith images are first considered. A second category of applications concern the analysis of 2D otolith growth. The potential of the proposed framework is illustrated on real otolith images for various species (e.g., cod, pollock) and discussed with a particular emphasis on the genericity of the approach and on applications such as otolith shape analysis, multi-proxy otolith analysis, otolith modeling.