We evaluate our method on two public cardiac MRI segmentation dataset. First the dataset from the Automated Cardiac Diagnosis Challenge (ACDC 2017) is employed. The second dataset used in this article is the M&M challenge dataset