Ischemic mitral regurgitation (IMR) is a disease where the normal mitral valve (MV) structure is dysfunctional due to left ventricular (LV) remodeling after a myocardial infarction (MI). IMR affects nearly 3 million Americans and the magnitude of this problem is expected to grow as the population ages. IMR has a substantial mortality rate that is associated with even mild MR severity. Mitral valve repair with undersized ring annuloplasty has been the preferred treatment strategy for IMR; however, the recurrence of moderate or severe IMR within 12 months of surgery is common. Recent results from the Cardiothoracic Surgical Trials Network (CTSN) multicenter randomized trials on IMR have confirmed a high incidence of early recurrent IMR. More importantly, these studies highlighted the adverse impact of recurrent IMR on LV remodeling and clinical outcomes. The CTSN trials demonstrated no significant difference in LV volume reduction or survival at 12 and 24 months between repair and replacement groups; however, subgroup analysis demonstrated that repair patients that developed recurrent IMR had no reduction in LV volume while repair patients without recurrence experienced LV volume reduction that was superior to patients having valve replacement. The results of the CTSN IMR trials indicate an unmet need for a pre-operative risk stratification tool that reliably predicts MV repair failure. Such a tool would significantly reduce the problem of recurrent IMR by performing valve repair only in patients likely to experience a durable result and performing valve replacement in patients with high risk of recurrence. The long term goal is to improve the quality of surgical therapy in IMR by improving risk-stratification pre-operatively using automated image analysis in the operating room. Current 3D computational algorithms have demonstrated success in creating reliable anatomical MV models from intraoperative 3D transesophageal echocardiography (3D-TEE), however they do not comprehensively assess IMR. While IMR does indeed manifest as MV malcoaptation, it is primarily caused by LV remodeling. The overall objective is to improve the prediction IMR recurrence by developing a fully automated left ventricle mitral valve (LVMV) model from 3D TEE. Towards this goal, the first part of the dissertation investigates the valvular-subvalvular interactions in IMR ovine models, demonstrates the clinical applicability of mitral valve modeling from 3D TEE and technically develops algorithms to generate patient-specific LVMV models. The second part of the dissertation demonstrates the potential applications outside of IMR such as modeling the tricuspid valve in hypoplastic left heart syndrome from echocardiography and investigates applications of relevant emerging technologies in the areas of machine learning and virtual reality.