Spatially Localized Atlas Network Tiles Enables 3D Whole Brain Segmentation from Limited Data

Huo, Yuankai; Xu, Zhoubing; Aboud, Katherine S.; Parvathaneni, Prasanna; Bao, Shunxing; Bermúdez, Camilo; Resnick, Susan M.; Cutting, Laurie E.; Landman, Bennett A.

doi:10.1007/978-3-030-00931-1_80

Cited by 28 publications

(24 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The study of this work is centered around a specific representative neuroscience application: SLANT [24,25]. This application performs multiple independent 3D fully convolutional network (FCN) for high-resolution whole brain segmentation.…”

Section: Spatially Localized Atlas Network Tiles (Slant)mentioning

confidence: 99%

Profiles of Upcoming HPC Applications and Their Impact on Reservation Strategies

Gainaru

Goglin

Honoré

et al. 2021

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

With the expected convergence between HPC, BigData and AI, new applications with different profiles are coming to HPC infrastructures. We aim at better understanding the features and needs of these applications in order to be able to run them efficiently on HPC platforms. The approach followed is bottom-up: we study thoroughly an emerging application, Spatially Localized Atlas Network Tiles (SLANT, originating from the neuroscience community) to understand its behavior. Based on these observations, we derive a generic, yet simple, application model (namely, a linear sequence of stochastic jobs). We expect this model to be representative for a large set of upcoming applications that require the computational power of HPC clusters without fitting the typical behavior of large-scale traditional applications. In a second step, we show how one can manipulate this generic model in a scheduling framework. Specifically we consider the problem of making reservations (both time and memory) for an execution on an HPC platform. We derive solutions using the model of the first step of this work. We experimentally show the robustness of the model, even with very few data or with another application, to generate the model, and provide performance gains with regards to standard and more recent approaches used in the neuroscience community.

show abstract

Section: Spatially Localized Atlas Network Tiles (Slant)mentioning

confidence: 99%

Profiles of Upcoming HPC Applications and Their Impact on Reservation Strategies

Gainaru

Goglin

Honoré

et al. 2021

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

show abstract

“…The original SLANT whole brain segmentation algorithm leverages a large dataset of automatically generated labels (N = 5,111 subjects) to pretrain the model before refining through TL using a small dataset of manually labelled ground truth atlases [5,6]. In this work, we begin with the pretrained model and use three different imaging datasets to test the effect of TL on algorithm generalizability.…”

Section: Imaging Datasetsmentioning

confidence: 99%

“…All models were trained with 80% of each of the subjects as the training set, while the remaining 20% was evenly split between validation and testing. Model accuracy was evaluated on the validation set using DSC, with the final model chosen as having the highest validation set DSC after 30 epochs of TL as in [5,6]. A five-fold cross-validation scheme was employed in independently trained models.…”

Section: Performance Analysismentioning

confidence: 99%

“…Until recently, a full-resolution, whole brain segmentation with over 100 labels was a challenging task due to hardware constraints. However, this problem was solved by Huo et al by using a tile-based method called the spatially localized atlas network tiles (SLANT) to divide T1-weighted brain MRI into 27 overlapping tiles and train independent networks, which are then fused via majority voting [5,6]. This method showed better performance than multi-atlas and other deep learning-based methods on brain MRI acquired for research.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Generalizing deep whole brain segmentation for pediatric and post-contrast MRI with augmented transfer learning

Bermúdez

Blaber

Remedios

et al. 2020

Medical Imaging 2020: Image Processing

Self Cite

View full text Add to dashboard Cite

Generalizability is an important problem in deep neural networks, especially in the context of the variability of data acquisition in clinical magnetic resonance imaging (MRI). Recently, the Spatially Localized Atlas Network Tiles (SLANT) approach has been shown to effectively segment whole brain non-contrast T1w MRI with 132 volumetric labels. Enhancing generalizability of SLANT would enable broader application of volumetric assessment in multi-site studies. Transfer learning (TL) is commonly used to update the neural network weights for local factors; yet, it is commonly recognized to risk degradation of performance on the original validation/test cohorts. Here, we explore TL by data augmentation to address these concerns in the context of adapting SLANT to anatomical variation and scanning protocol. We consider two datasets: First, we optimize for age with 30 T1w MRI of young children with manually corrected volumetric labels, and accuracy of automated segmentation defined relative to the manually provided truth. Second, we optimize for acquisition with 36 paired datasets of pre-and post-contrast clinically acquired T1w MRI, and accuracy of the post-contrast segmentations assessed relative to the pre-contrast automated assessment. For both studies, we augment the original TL step of SLANT with either only the new data or with both original and new data. Over baseline SLANT, both approaches yielded significantly improved performance (signed rank tests; pediatric: 0.89 vs. 0.82 DSC, p<0.001; contrast: 0.80 vs 0.76, p<0.001). The performance on the original test set decreased with the new-data only transfer learning approach, so data augmentation was superior to strict transfer learning.

show abstract

“…> 30 hours per scan) restricts the MAS whole brain segmentation on large-scale clinical cohorts. Recent advances in deep convolutional neural networks (DCNN) provide powerful tools to alleviate the computational time for whole brain segmentation [4][5][6][7][8], and even may result in higher accuracy compared with MAS such as the recently proposed spatially localized atlas network tiles (SLANT) method [9] (https://github.com/MASILab/SLANT_brain_seg). The SLANT method has been shown to achieve superior performance on small-scale validation methods as it was trained on >5000 MRI scans from >60 sites [10].…”

Section: Introductionmentioning

confidence: 99%

Reproducibility evaluation of SLANT whole brain segmentation across clinical magnetic resonance imaging protocols

Xiong

Huo

Wang

et al. 2019

Medical Imaging 2019: Image Processing

Self Cite

View full text Add to dashboard Cite

Whole brain segmentation on structural magnetic resonance imaging (MRI) is essential for understanding neuroanatomical-functional relationships. Traditionally, multi-atlas segmentation has been regarded as the standard method for whole brain segmentation. In past few years, deep convolutional neural network (DCNN) segmentation methods have demonstrated their advantages in both accuracy and computational efficiency. Recently, we proposed the spatially localized atlas network tiles (SLANT) method, which is able to segment a 3D MRI brain scan into 132 anatomical regions. Commonly, DCNN segmentation methods yield inferior performance under external validations, especially when the testing patterns were not presented in the training cohorts. Recently, we obtained a clinically acquired, multi-sequence MRI brain cohort with 1480 clinically acquired, de-identified brain MRI scans on 395 patients using seven different MRI protocols. Moreover, each subject has at least two scans from different MRI protocols. Herein, we assess the SLANT method’s intra- and inter-protocol reproducibility. SLANT achieved less than 0.05 coefficient of variation (CV) for intra-protocol experiments and less than 0.15 CV for inter-protocol experiments. The results show that the SLANT method achieved high intra- and inter- protocol reproducibility.

show abstract

Spatially Localized Atlas Network Tiles Enables 3D Whole Brain Segmentation from Limited Data

Cited by 28 publications

References 15 publications

Profiles of Upcoming HPC Applications and Their Impact on Reservation Strategies

Profiles of Upcoming HPC Applications and Their Impact on Reservation Strategies

Generalizing deep whole brain segmentation for pediatric and post-contrast MRI with augmented transfer learning

Reproducibility evaluation of SLANT whole brain segmentation across clinical magnetic resonance imaging protocols

Contact Info

Product

Resources

About