Cosmic voids are effective cosmological probes to discriminate among competing world models. Their identification is generally based on density or geometry criteria that, because of their very nature, are prone to shot noise. We propose two void finders that are based on dynamical criterion to select voids in Lagrangian coordinates and minimise the impact of sparse sampling. The first approach exploits the Zel'dovich approximation to trace back in time the orbits of galaxies located in voids and their surroundings, the second uses the observed galaxy-galaxy correlation function to relax the objects' spatial distribution to homogeneity and isotropy. In both cases voids are defined as regions of the negative velocity divergence, that can be regarded as sinks of the back-in-time streamlines of the mass tracers. To assess the performance of our methods we used a dark matter halo mock catalogue CoDECS, and compared the results with those obtained with the ZOBOV void finder. We find that the void divergence profiles are less scattered than the density ones and, therefore, their stacking constitutes a more accurate cosmological probe. The significance of the divergence signal in the central part of voids obtained from both our finders is 60% higher than for overdensity profiles in the ZOBOV case. The ellipticity of the stacked void measured in the divergence field is closer to unity, as expected, than what is found when using halo positions. Therefore our void finders are complementary to the existing methods, that should contribute to improve the accuracy of void-based cosmological tests.