Jiri Hladuvka

MDL Spline Models: Gradient and Polynomial Reparameterisations.

The major problem when building shape models by minimising the description length (DL) is the computationally demanding optimisation in a highdimensional space. To speed up the convergence when dealing with discretised contours, Ericsson and Astrom [6] have shown how to approximate and how to exploit the gradient of the DL. We derive the gradient of the DL for differentiable training sets. Additionally, we propose a class of polynomial reparameterisations that allows us to avoid numerical approximation of the functions and integrals involved in computation of the model and of its DL’s gradient, making the whole process exact and more efficient.

Automated ROI placement and trabecula-driven orientation for radiographic texture analyses of calcaneus

Radiographic texture analysis of calcaneus has been recently extensively researched for automated discrimination between osteoporotic and healthy subjects. Existing work concentrates on design of texture descriptors and assumes already available texture patches previously defined by human experts. In this paper we address the missing link to complete the automatic workflow — localization and orientation of regions of interest (ROIs) the texture patches are sampled from. We exploit the shape of calcaneus to define ROI position and the structure of trabecula to define its orientation. We evaluated our automated method on 361 high-resolution radiographs. Comparison to user-defined ROIs reveals localization discrepancy of 1.85 ± 1.06 mm and deviation in orientation of −1.81 ± 7.31 degrees.