A new framework for morphological and morphometric study of fish species based on groupwise registration of otolith images

Morphology of bones, teeth, and some particular structures are widely used for categorizing species and studying their evolution. In this paper, we used groupwise registration to provide a representative image from the set of the image samples that represents its typical morphology. We also provided perturbation map which indicates the deviation of each point through the ensemble. These images support qualitative discussions about the morphology of structures in different species. The perturbation map can be further exploited for determining appropriate landmarks for morphometric analysis. Knowing the deformation between the prototype and each image sample from the species, the framework allows for automatic detection of corresponding points. Once the user puts a landmark on the prototype image, the corresponding points on all the image samples will be determined. This maximizes the accuracy in measuring the morphological indices by eliminating the human error due to uncertainty in locating landmarks.