Adaptive Bayesian label fusion using kernel-based similarity metrics in hippocampus segmentation

Abstract

The effectiveness of brain magnetic resonance imaging (MRI) as a useful evaluation tool strongly depends on the performed segmentation of associated tissues or anatomical structures. We introduce an enhanced brain segmentation approach of Bayesian label fusion that includes the construction of adaptive target-specific probabilistic priors using atlases ranked by kernel-based similarity metrics to deal with the anatomical variability of collected MRI data. In particular, the developed segmentation approach appraises patch-based voxel representation to enhance the voxel embedding in spaces with increased tissue discrimination, as well as the construction of a neighborhood-dependent model that addresses the label assignment of each region with a different patch complexity. To measure the similarity between the target and training atlases, we propose a tensor-based kernel metric that also includes the training labeling set. We evaluate the proposed approach, adaptive Bayesian label fusion using kernel-based similarity metrics, in the specific case of hippocampus segmentation of five benchmark MRI collections, including ADNI dataset, resulting in an increased performance (assessed through the Dice index) as compared to other recent works.

Feedback Chain Network For Hippocampus Segmentation

“…[45] as comparisons to enhance the validation of our proposed algorithm. For the IBSR dataset, the methods compared add Rousseau's method [35], Zarpalas's method [36], David's method [37] and Shi's method [45]. Among them, the implementation codes of the methods are not open source except Thyreau et al [39] and Shi et al [45], so the performance scores are taken from the corresponding papers.…”

“…Before deep neural networks were widely applied to medical image segmentation, most of the best automatic hippocampal segmentation methods used multi-atlas approaches [34]- [37]. Until today, the method proposed by FreeSurfer [38] is still followed for segmentation of brain structures in medical research.…”

Entropy-Based Estimation of Event-Related De/Synchronization in Motor Imagery Using Vector-Quantized Patterns

“…Assuming that the input samples follow a Gaussian distribution, we employ the similarity-based approach between sets for estimation of both probabilistic terms, as proposed in [ 31 ]: where is the probability that a symbol belongs to every element of the dictionary, , being a Gaussian similarity function, and the moments computed, respectively, as below: …”