 We developed a novel method for mapping between time-varying cardiac geometries obtained using different acquisition and analysis protocols. This enabled us to reduce measurement biases specific to each protocol and enable the pooling of data between modalities. The method was demonstrated on paired real-time 3D echocardiography, 3D and cardiac magnetic resonance, CMR, sequences from 138 subjects. After applying our method, we observed a significant reduction in mean bias, narrower limits of agreement, and higher intra-class correlation coefficients for all functional indices between CMR and 3D geometries. Average root mean squared errors between surface coordinates of 3D and CMR geometries also decreased from 7 plus or minus 1 to 4 plus or minus 1 millimeter for the total study population. This article was authored by Debbie Zhao, Charlene Morgan, Kathleen Gilbert, and others.