Asha and Keshav Bhide Chair Professor
Computer Science and Engineering Department, Indian Institute of Technology (IIT) Bombay
Office: A-214, Kanwal Rekhi Building
Email: my-first-name @cse.iitb.ac.in
|
|
Suyash P.
Awate Asha and Keshav Bhide Chair Professor Computer Science and Engineering Department, Indian Institute of Technology (IIT) Bombay Office: A-214, Kanwal Rekhi Building Email: my-first-name @cse.iitb.ac.in |
| Research | Publications | Teaching | Students | CV | Personal |
| Multiatlas Segmentation as Nonparametric Regression |
| Suyash P. Awate, Peihong Zhu, Ross T. Whitaker How many templates does it take for a good segmentation?: error analysis in multiatlas segmentation as a function of database size Int. Workshop Multimodal Brain Image Analysis (MBIA) at Int. Conf. MICCAI 2012, 2:103-114, Springer LNCS 7509 (podium presentation, Best Paper Award) Suyash P. Awate, Ross T. Whitaker Multiatlas Segmentation as Nonparametric Regression IEEE Transactions on Medical Imaging (TMI) 33(9): 1803-1817 (2014)  Predicting MSE (for fixed k) as a function of database size for (a) subcortical Structures in brain MR images and (b) cartilages in knee MR Images (see paper). Box-whisker plots show the empirically computed MSE values after size normalization for the number of neighbors/templates in kNN fixed to eight independent of database size. Variance depicted by the whiskers comes from the bootstrap samples of small databases available for analysis. Solid lines and associated error bars are obtained using the larger databases. Dashed lines and associated error bars show the mean and standard deviation, respectively, of the fitted MSE values using small databases . Error bars come from the bootstrap samples of databases.  Parameter values for multiatlas caudate segmentation from brain T1 MR images using small databases, indicating confidence of labels per voxel. Images (a) and (e) show the average MR image and the average expert segmentation, respectively, in a common anatomical space. Images (b) and (f) show the mean and standard deviation, respectively, of alpha_v, at voxel v, generated from bootstrap samples of the small databases employed for parameter estimation. Similarly, (c) and (g) show the mean and standard deviation of and (d) and (h) show the mean and standard deviation of d_v. |
| Related Works |
| Juan Eugenio Iglesias, Mert R. Sabuncu Multi-atlas segmentation of biomedical images: A survey Medical Image Analysis 24(1): 205-219 (2015) |