A decomposition scheme for 3D fuzzy objects is presented. The decomposition is guided by a fuzzy distance transform (FDT) of the fuzzy object and aims to decompose the fuzzy object into simpler parts. Relevant voxels, corresponding to the ‘‘centres’’ of the parts, are detected on the FDT and suitably grouped, using a hierarchical clustering technique, into significant seeds for the decomposition. A region growing process is then applied to the seeds. The region growing process makes use of the reverse fuzzy distance transform, which is introduced in this manuscript. The decomposition scheme is illustrated using real data from different applications of which one, namely the identification of the three parts of the Immunoglobulin G antibody imaged using cryo electron tomography, is described more in detail. Ó 2006 Elsevier B.V. All rights reserved.