The clinical and biomechanical communities are increasingly interested in obtaining accurate and reliable three-dimensional reconstructions of the motion of the heart chambers from routine time-resolved medical images, such as cardiac magnetic resonance imaging (MRI) or computed tomography (CT). This methodology plays a crucial role in enhancing diagnostic tools based on regional deformation patterns, isolating regions of interest in four-dimensional phase-contrast (4D Flow) MRI datasets, or providing boundary conditions for image-based biomechanical simulations, among other applications. This study critically reviews and assesses both standard and novel techniques for the reconstruction of the heart kinematics, including a recently-developed method based on a combination of feature-tracking algorithms and diffeomorphic mappings for standard multi-slice cardiac MRI [1]. The emphasis is particularly on the right ventricle (RV) of the heart, given the complexity of RV shape and the urgent clinical need to enhance tools for characterizing RV function in health and disease. Importantly, the work focuses on the critical issue of assessing the accuracy and reliability of the reconstructed motion. Traditionally, this assessment is carried out in a static way, by comparing the reconstructed shape with the ground truth at each time frame using image similarity measures. In this work, we underscore the significance of a dynamic assessment against flow-related measures, i.e. checking the physical compatibility of the reconstructed motion in terms of i) mass conservation, and ii) hemodynamic forces. The concept is applied to various cases of both healthy and diseased hearts.