Autism is a developmental disorder characterized by social deficits, impaired communication, and restricted and repetitive patterns of behaviour. Emerging theories indicate interregional functional and anatomical brain connectivity as a likely key feature in autism pathophysiology. Corpus callosum (CC) represents a natural target of autism connectivity research, being the expression of interhemispheric communication. In this paper, a novel method for a robust morphometric analysis of CC data is presented. The standard morphometric approach is based on the analysis of the size and shape of the CC midsagittal cross-section. As there are no gross anatomical landmarks that clearly delimit anatomically or functionally distinct CC regions, several geometric partitioning schemes have been proposed in the literature for morphometric analysis, subdividing CC into subregions whose fiber topography is expected to target different hemispheric cortical regions. A novel tool of morphometric analysis, based on the automated subdivision of a high number of partitions from a CC centroid and on the consequent determination of the CC anatomical landmarks is presented, allowing an automated analysis of CC volumes, shapes and curvatures, suitable for an automated application in clinical environment. Moreover the proposed tool can be used for original post-processing and visualization techniques that may help in the analysis of possible alterations of CC and in the correlations with autism-related diseases. The proposed morphometric tool has been validated and applied for clinical investigation on brain morphometry in children (age 3-11 years) with autism or with other autism spectrum disorders (DSA) and on healthy control subjects who underwent volumetric MRI T1 weighted acquisitions.