Ventricular (VT) and atrial (AT) tachycardias are some of the most common clinical cardiac arrhythmias. For ablation of tachycardia substrates, two clinical diagnosis methods are used: invasive electroanatomical mapping for an accurate diagnosis using electrograms (EGMs) acquired with intracardiac catheters, and localized on the surface mesh of the studied cavities; and noninvasive electrocardiographic imaging (ECGi) for a global view of the arrhythmia, with EGMs mathematically reconstructed from body surface electrocardiograms using 3-D cardio-thoracic surface meshes obtained from CT-scans. In clinics, VT and AT are diagnosed by studying activation time maps that depict the propagation of the activation wavefront on the cardiac mesh. Nevertheless, slow conduction areas-a well-known proarrhythmic feature for tachycardias-and tachycardia specific propagation patterns are not easily identifiable with these maps. Therefore, local characterization of the activation wavefront propagation can be helpful for improving VT and AT diagnoses. The purpose of this study is to develop a method to locally characterize the activation wavefront propagation for clinical data. For this, a conduction velocity vector field is estimated and analyzed using divergence and curl mathematical operators. The workflow was first validated on a simulated database from computer models, and then applied to a clinical database obtained from ECGi to improve AT diagnosis. The results show the relevancy and the efficacy of the proposed method to guide ablation of tachyarrhythmias.< Réduire