Maximum-Likelihood estimation of biological growth variables

Authors:
Anuj Srivastava;Sanjay Saini;Zhaohua Ding;Ulf Grenander
Affiliations:
Department of Statistics, Florida State University, Tallahassee, FL;Department of Statistics, Florida State University, Tallahassee, FL;Institute of Imaging Sciences, Vanderbilt University, Nashville, TN;Division of Applied Mathematics, Brown University, Providence, RI
Venue:
EMMCVPR'05 Proceedings of the 5th international conference on Energy Minimization Methods in Computer Vision and Pattern Recognition
Year:
2005

Citing 4
Cited 1

Diffeomorphisms Groups and Pattern Matching in Image Analysis

International Journal of Computer Vision
Dynamic Programming Generation of Curves on Brain Surfaces

IEEE Transactions on Pattern Analysis and Machine Intelligence
Computational anatomy: an emerging discipline

Quarterly of Applied Mathematics - Special issue on current and future challenges in the applications of mathematics
Group Actions, Homeomorphisms, and Matching: A General Framework

International Journal of Computer Vision - Special issue on statistical and computational theories of vision: Part II

Direct Estimation of Biological Growth Properties from Image Data Using the "GRID" Model

ICIAR '09 Proceedings of the 6th International Conference on Image Analysis and Recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

Shapes of biological objects, such as anatomical parts, have been studied intensely in recent years. An emerging need is to model and analyze changes in shapes of biological objects during, for example, growths of organisms. A recent paper by Grenander et al. [5] introduced a mathematical model, called GRID, for decomposing growth induced diffeomorphism into smaller, local deformations. The basic idea is to place focal points of local growth, called seeds, according to a spatial process on a time-varying coordinate system, and to deform a small neighborhood around them using radial deformation functions (RDFs). In order to estimate these variables – seed placements and RDFS – we first estimate optimal deformation from magnetic resonance image data, and then utilize an iterative solution to reach maximum-likelihood estimates. We demonstrate this approach using MRI images of human brain growth.