Adrian Vasile Dalca

Postdoctoral Fellow     CV | Linkedin

Computer Science and Artificial Intelligence Lab
EECS, Massachusetts Institute of Technology

A.A. Martinos Center for Biomedical Imaging
Massachusetts General Hospital, Harvard Medical School

Contact:
32 Vassar St, 32-G904, Cambridge, MA, 02139 
adalca at mit dot edu

Research

My research focuses on probabilistic mathematical models and machine learning techniques that capture important relationships between medical images, clinical diagnoses, and other rich biomedical data.

This page describes my current main directions, although I'm often interested in a variety of subjects, spanning subjects like computer vision, physics, and machine learning for health — as seen on my publications list.

I am fortunate to work with exceptional collaborators Polina Golland, Mert Sabuncu, John Guttag, Natalia Rost, Katie Bouman, Ramesh Sridharan, Danielle Pace, Guha Balakrishnan, Amy Zhao, Andreea Bobu, Kayhan Batmanghelich, Ehud Schmidt, Manolis Kellis and Jonathan Rosand. I've also done interesting side projects with David Gifford. I've been very lucky to work in bioinformatics with professor Michael Brudno and in geophysics with professor Jerry Mitrovica at the University of Toronto.

Current Research

Population structure for subject-specific inference

A main focus is developing analytic techniques that learn concise representations from collections of biomedical images and enable novel insight about the anatomy and pathology of a new patient.

For example, to facilitate the use of sparse clinical scans for scientific study, we define image imputation, the statistical inference of unobserved anatomical slices using a collection of low resolution volumes. We similarly use population structure and anatomical and disease priors to propose novel image segmentation algorithms.


  • A.V. Dalca, K.L. Bouman, W.T. Freeman, N.S. Rost, M.R. Sabuncu, and P. Golland. Medical Image Imputation from Image Collections Under Review at IEEE: Transactions on Medical Imaging

  • A.V. Dalca, J. Guttag, and M. R. Sabuncu.
    Anatomical Priors in Convolutional Networks for Unsupervised Biomedical Segmentation.
    In Proc. CVPR: Conference on Computer Vision and Pattern Recognition. 2018
  • A.V. Dalca, J. Guttag, and M. R. Sabuncu.
    Spatial Anatomical Priors in Convolutional Networks for Unsupervised Biomedical Segmentation
    NIPS ML4H: Machine Learning for Health. 2017. Spotlight
  • A.V. Dalca, K.L. Bouman, W.T. Freeman, M.R. Sabuncu, N.S. Rost, P. Golland.
    Population Based Image Imputation.
    In Proc. IPMI: International Conference on Information Processing and Medical Imaging. LNCS 10265, pp 659-671. 2017. Best poster award
  • A.V. Dalca, R. Sridharan, L. Cloonan, K. M. Fitzpatrick, A. Kanakis, K.L. Furie, J.Rosand, O.Wu, M.Sabuncu, N.S. Rost, P.Golland.
    Segmentation of Cerebrovascular Pathologies in Stroke Patients with Spatial and Shape Priors.
    In Proc. MICCAI: International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), 2014.
  • Abstract at MICGen 2014.
  • R. Sridharan, A.V. Dalca, K.M. Fitzpatrick, L. Cloonan, A. Kanakis, O. Wu, K.L. Furie, J. Rosand, N.S. Rost, P. Golland.
    Quantification and Analysis of Large Multimodal Clinical Image Studies: Application to Stroke.
    In Proc. MICCAI International Workshop on Multimodal Brain Image Analysis (MBIA), pp. 18–30, 2013.
    equal contribution

Biomedical data and anatomy

Anatomical variability is closely related to factors external to the medical scan. I am interested in how external data, like patient genotypes, can help anatomical analysis of medical images, and I actively organize the MICGen workshop in Imaging Genetics to bring together research on these topics. Integrating and characterizing imaging together with other medical data will establish a holistic understanding of not just a patient's current health, but his or her future trajectory of anatomical changes and health.

For example, we explored this direction starting with a simple question: can genetic variants predict (aspects of) entire MR scans? We developed a semi-parametric generative Gaussian Process model for predicting the anatomy of a patient in subsequent brain scans following a baseline scan.

  • Co-organizing the third MICGen: MICCAI Workshop on Imaging Genetics at MICCAI 2017.
  • A.V. Dalca, N.K. Batmanghelich, M.R. Sabuncu, L. Shen (Editors). Imaging Genetics. Elsevier. 2017.
  • K.N. Batmanghelich, A.V. Dalca, G. Quon, M.R. Sabuncu, P. Golland. Probabilistic Modeling of Imaging, Genetics and Diagnosis. IEEE Transactions on Medical Imaging, 35(7), pp 1765-79, 2016.
  • A.V. Dalca, R. Sridharan, M.R. Sabuncu, P. Golland. Predictive Modeling of Anatomy with Genetic and Clinical Data. In Proc. MICCAI: International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), LNCS 9351, pp. 519-526, 2015.
  • K.N. Batmanghelich, A.V. Dalca, M.R. Sabuncu, P. Golland. Joint Modeling of Imaging and Genetics, In Proc. IPMI: International Conference on Information Processing and Medical Imaging, LNCS 7917, pp. 766-777, 2013.

Clinical Translation

I am interested in developing tools to enable productive collaboration with clinicians to help impact clinical research and treatment. The work often draws on technical insights from several fields. For example, we've developed methods for image registration and visualization that explicitly tackle difficulties in clinical images. We also model disease progression using a nonparametric regression mixture model, providing examination of spatial progression of cerebrovascular disease with age in thousands of patients.


  • A.V. Dalca, G. Balakrishnan, J. Guttag, and M.R. Sabuncu.
    Unsupervised Learning for Fast Probabilistic Diffeomorphic Registration.
    MICCAI: International Conference on Medical Image Computing and Computer Assisted Intervention 2018
    Early Accept
  • G. Balakrishnan, A. Zhao, M. R. Sabuncu, J. Guttag, and A.V. Dalca.
    An Unsupervised Learning Model for Deformable Medical Image Registration.
    In Proc. CVPR: Conference on Computer Vision and Pattern Recognition. 2018
  • A.K. Giese, M.D. Schirmer, K.L. Donahue, L. Cloonan, R. Irie, S. Winzeck, M.J.R.J. Bouts, E.C. McIntosh, S. J. Mocking, A.V. Dalca, et al. Design and Rationale for Examining Neuroimaging Genetics in Ischemic Stroke: the MRI-GENIE Study. Neurology: Genetics. 3(5). 2017.
  • A.V. Dalca, A. Bobu, N.S. Rost, P. Golland. Patch-Based Discrete Registration of Clinical Brain Images. In Proc. MICCAI-PATCHMI Patch-based Techniques in Medical Imaging, LNCS 9993, pp 60-67, 2016. Best paper award
  • A.V. Dalca, R. Sridharan, N.S. Rost, P. Golland. tipiX: Rapid Visualization of Large Image Collections In MICCAI-IMIC Interactive Medical Image Computing Workshop, 2014. Best paper award for impact and usability.; Finalist CSAIL Amazing Research Highlight Competition.
  • R. Sridharan, A.V. Dalca, P. Golland. An interactive visualization tool for Nipype medical imaging pipelines. In MICCAI-IMIC Interactive Medical Image Computing Workshop, 2014.
  • R. Sridharan, A.V. Dalca, K.M. Fitzpatrick, L. Cloonan, A. Kanakis, O. Wu, K.L. Furie, J. Rosand, N.S. Rost, P. Golland. Quantification and Analysis of Large Multimodal Clinical Image Studies: Application to Stroke. In Proc. MICCAI International Workshop on Multimodal Brain Image Analysis (MBIA), pp. 18–30, 2013.
    equal contribution

Webdesign: Adrian Dalca. Based on: MiniFolio