This report describes an algorithm for the generation of unstructured tetrahedral meshes from two triangular meshes: the internal and external walls of an object. The tetrahedral mesh thus obtained represents the volume between the two given triangular meshes. We choose tetrahedral meshes to represent the volume because they provide more topological flexibility. The algorithm designed for this purpose is based on advancing front techniques, but unlike most existing methods based on this concept, our algorithm uses adjacency information about the elements of the volumetric mesh to construct a tetrahedral element. In this manner, it is possible to select, at each step, the neighboring point that guarantees the best tetrahedron, based on an objective function. In addition, the proposed algorithm allows us to have, within the same model, morphological and functional information of interest. This information is very useful, for instance, in the reconstruction of the left ventricle (LV) of the heart, as will be discussed in the paper.