Automatic segmentation of the heart cavity is an essential task for the diagnosis of cardiac diseases. In this paper, we propose a semi-supervised segmentation setup for leveraging unlabeled data to segment Left-ventricle, Right-ventricle, and Myocardium. We utilize an enhanced version of residual U-Net architecture on a large-scale cardiac MRI dataset. Handling the class imbalanced data issue using dice loss, the enhanced supervised model is able to achieve better dice scores in comparison with a vanilla U-Net model. We applied several augmentation techniques including histogram matching to increase the performance of our model in other domains. Also, we introduce a simple but efficient semi-supervised segmentation method to improve segmentation results without the need for large labeled data. Finally, we applied our method on two benchmark datasets, STACOM2018, and M\&Ms 2020 challenges, to show the potency of the proposed model. The effectiveness of our proposed model is demonstrated by the quantitative results. The model achieves average dice scores of 0.921, 0.926, and 0.891 for Left-ventricle, Right-ventricle, and Myocardium respectively.