Combining Registration Errors and Supervoxel Classification for Unsupervised Brain Anomaly Detection

Martins, Samuel Botter; Falcão, Alexandre X.; Telea, Alexandru

doi:10.1007/978-3-030-72379-8_8

Cited by 2 publications

(2 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The reach of the ATLAS v1.2 dataset has also extended beyond stroke lesion segmentation. It has also been used as a key example of a large, public neuroimaging dataset, 29 to provide published guidelines on how to perform lesion segmentation, 30 to evaluate the performance of different hippocampal segmentation methods in stroke, 31 to test other non-stroke automated methods, such as anomaly 32 and asymmetry detection, 33 and as inspiration for future AI programs and large public datasets, 34 among other uses. It is a valuable educational resource and has been used as a teaching resource in courses on machine learning and computer vision as well as for student thesis projects.…”

Section: Background and Summarymentioning

confidence: 99%

A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms

Liew

Donnelly

et al. 2021

Preprint

View full text Add to dashboard Cite

Accurate lesion segmentation is critical in stroke rehabilitation research for the quantification of lesion burden and accurate image processing. Current automated lesion segmentation methods for T1-weighted (T1w) MRIs, commonly used in rehabilitation research, lack accuracy and reliability. Manual segmentation remains the gold standard, but it is time-consuming, subjective, and requires significant neuroanatomical expertise. We previously released a large, open-source dataset of stroke T1w MRIs and manually segmented lesion masks (ATLAS v1.2, N=304) to encourage the development of better algorithms. However, many methods developed with ATLAS v1.2 report low accuracy, are not publicly accessible or are improperly validated, limiting their utility to the field. Here we present ATLAS v2.0 (N=955), a larger dataset of T1w stroke MRIs and manually segmented lesion masks that includes both training (public) and test (hidden) data. Algorithm development using this larger sample should lead to more robust solutions, and the hidden test data allows for unbiased performance evaluation via segmentation challenges. We anticipate that ATLAS v2.0 will lead to improved algorithms, facilitating large-scale stroke rehabilitation research.

show abstract

Section: Background and Summarymentioning

confidence: 99%

A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms

Liew

Donnelly

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The thesis is the result of the following publications (in chronological order): [32], [33] (honorable mention), [9], [10], [17], [22], [28], [8] (best student paper), [34].…”

Section: Acknowledgmentmentioning

confidence: 99%

Unsupervised Brain Anomaly Detection in MR Images

Martins

Telea

Falcão

2021

Anais Estendidos Da XXXIV Conference on Graphics, Patterns and Images (SIBRAPI Estendido 2021)

View full text Add to dashboard Cite

Many brain anomalies are associated with abnormal asymmetries. To detect and/or segment such anomalies in brain images, most automatic methods rely on supervised learning. This requires a large number of high-quality annotated training images, which is lacking for most medical image analysis problems. In contrast, unsupervised methods aim to learn a model from unlabeled healthy images, so that an unseen image that breaks priors of this model, i.e., an outlier, is considered an anomaly. This paper addresses the development of solutions to leverage unsupervised machine learning for the detection/analysis of abnormal brain asymmetries related to anomalies in magnetic resonance (MR) images. Experimental results on 3D MR-T1 images from healthy subjects and patients with a variety of lesions show the effectiveness and robustness of the proposed unsupervised approaches for brain anomaly detection.

show abstract

Combining Registration Errors and Supervoxel Classification for Unsupervised Brain Anomaly Detection

Cited by 2 publications

References 40 publications

A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms

A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms

Unsupervised Brain Anomaly Detection in MR Images

Contact Info

Product

Resources

About