Segmentation of brain tumors in 4D MR images using the hidden Markov model

Authors:
Jeffrey Solomon;John A. Butman;Arun Sood
Affiliations:
Medical Numerics Inc., Sterling, VA, USA;Department of Radiology, National Institutes of Health, Bethesda, MD, USA;Center for Image Analysis, George Mason University, Fairfax, VA, USA
Venue:
Computer Methods and Programs in Biomedicine
Year:
2006

Citing 3
Cited 5

Spatio-temporal Segmentation of Active Multiple Sclerosis Lesions in Serial MRI Data

IPMI '01 Proceedings of the 17th International Conference on Information Processing in Medical Imaging
Automatic Segmentation of Brain Tissues and MR Bias Field Correction Using a Cigital Brain Atlas

MICCAI '98 Proceedings of the First International Conference on Medical Image Computing and Computer-Assisted Intervention
Artificial Intelligence: A Modern Approach

Artificial Intelligence: A Modern Approach

Editorial: Medical image segmentation: Quo Vadis

Computer Methods and Programs in Biomedicine
3D medical volume segmentation using hybrid multiresolution statistical approaches

Advances in Artificial Intelligence - Special issue on machine learning paradigms for modeling spatial and temporal information in multimedia data mining
Adaptive pixon represented segmentation (APRS) for 3D MR brain images based on mean shift and Markov random fields

Pattern Recognition Letters
Neuro-levelset system based segmentation in dynamic susceptibility contrast enhanced and diffusion weighted magnetic resonance images

Pattern Recognition
Segmentation of brain tissues using a 3-D multi-layer Hidden Markov Model

Computers in Biology and Medicine

Quantified Score

Hi-index	0.00

Visualization

Abstract

Tumor size is an objective measure that is used to evaluate the effectiveness of anticancer agents. Responses to therapy are categorized as complete response, partial response, stable disease and progressive disease. Implicit in this scheme is the change in the tumor over time; however, most tumor segmentation algorithms do not use temporal information. Here we introduce an automated method using probabilistic reasoning over both space and time to segment brain tumors from 4D spatio-temporal MRI data. The 3D expectation-maximization method is extended using the hidden Markov model to infer tumor classification based on previous and subsequent segmentation results. Spatial coherence via a Markov Random Field was included in the 3D spatial model. Simulated images as well as patient images from three independent sources were used to validate this method. The sensitivity and specificity of tumor segmentation using this spatio-temporal model is improved over commonly used spatial or temporal models alone.