Omnipolar electrograms (oEGMs) provide orientation-independent characterisation of cardiac electrical activity by estimating the spatial gradient of electrical potential within small groups of adjacent electrodes called cliques. Their accuracy deteriorates when the locally planar wavefront assumption is violated — a condition frequent in fibrotic tissue and at the 4 mm inter-electrode spacings of current clinical catheters. We propose an interpolation-based method that reduces the effective inter-electrode distance to 0.2 mm using Gaussian radial basis function (RBF) interpolation, computing oEGMs at four virtual cross-shaped electrodes placed near each clique centroid. Three catheter geometries (polar, staggered, HD Grid) with N = 4–25 electrodes were evaluated on in silico anisotropic and fibrotic cardiac tissue. Performance was assessed via Pearson correlation, LAT error, and the Residuum-to-Omnipolar Ratio (ROR). Results showed a poor performance from the polar configuration and was discarded due to consistently poor performance. A staggered 5×5 array (N = 25) outperformed the commercially available Abbott HD Grid 4×4 (N = 16) across all metrics. The interpolated method consistently surpassed the traditional approach; restored tissue discriminability (ROR gap: 0.044 → 0.202), representing the most clinically relevant improvement.