creator Parages, Felipe M. advisor Brankov, Jovan Estimation of myocardial motion from medical images is an important problem in several fields such as computer-aided diagnosis of cardiac lesions, development of cardiac simulators, and cardiac image processing in general. Tagged Magnetic Resonance Imaging (tagged MRI) is an image modality in which magnetic markers (taglines) track true deformation of heart tissue, thus being a valuable tool to analyze cardiac motion. Although several methods to analyze cardiac motion from tagged MRI have been proposed in the past, many of them rely somewhat on manual intervention to correct loss of tagline contrast towards the end of the cardiac cycle. Likewise, some of these methods are based on tagged MRI only, failing to include motion information from untagged (cine) gated cardiac MRI. We present in this thesis a model that overcomes these limitations by using a regularized non-rigid Deformable Mesh Model (DMM) to track pixel-intensity variations over combined datasets of gated tagged and cine MRI. DMM performs motion estimation directly over MRI pixel intensities, and proved to achieve motion analysis results comparable to other methods in which manual fine-tuning of tagline tracking is required. In addition, incorporating information from gated cine MRI into DMM proves to be helpful into capturing cardiac motion more efficiently. Finally, we show in this work the capacity of DMM to accurately estimate geometric and mechanical properties of the myocardium, such as strain and ejection fraction. We conclude that DMM method has a great potential for cardiac motion modeling, and may constitute a good alternative to other techniques in the future. Submitted by Dana Lamparello (dlampare@iit.edu) on 2012-02-29T17:14:26Z No. of bitstreams: 2 ParagesMSthesis.pdf: 2057872 bytes, checksum: 671b589eed9d4221b2d07fb05c43ac25 (MD5) ParagesMScover.pdf: 207484 bytes, checksum: 8554e70a0172510af253587eb5a56d5c (MD5) Made available in DSpace on 2012-02-29T17:14:26Z (GMT). No. of bitstreams: 2 ParagesMSthesis.pdf: 2057872 bytes, checksum: 671b589eed9d4221b2d07fb05c43ac25 (MD5) ParagesMScover.pdf: 207484 bytes, checksum: 8554e70a0172510af253587eb5a56d5c (MD5) Previous issue date: 2011-05 M.S. in Electrical Engineering, May 2011 2011-04-25 2011-05 http://hdl.handle.net/10560/2533 en Thesis born digital application/pdf In Copyright http://rightsstatements.org/page/InC/1.0/ Restricted Access ECE / Electrical and Computer Engineering Illinois Institute of Technology Affiliated department