Automatic rat brain segmentation from MRI using statistical shape models and random forest

Bendazzoli, Simone; Brusini, Irene; Damberg, Peter; Smedby, Örjan; Andersson, Leif; Wang, Chunliang

doi:10.1117/12.2512409

Cited by 8 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To date, several attempts have been made to address rodent skull-segmentation ( Pfefferbaum et al, 2004 ; Sharief et al, 2008 ; Bendazzoli et al, 2019 ; Feo and Giove, 2019 ; Lohmeier et al, 2019 ; Liu et al, 2020 ). To date, the most prominent tools for rodent MRI skull stripping are Pulse-Coupled Neural Network (PCNN)-based brain extraction proposed by Chou et al (2011) , Rapid Automatic Tissue Segmentation (RATS) pioneered by Oguz et al (2014) , and, and SHape descriptor selected External Regions after Morphologically filtering (SHERM) by Liu et al (2020) .…”

Section: Introductionmentioning

confidence: 99%

Automatic Skull Stripping of Rat and Mouse Brain MRI Data Using U-Net

et al. 2020

View full text Add to dashboard Cite

Accurate removal of magnetic resonance imaging (MRI) signal outside the brain, a.k.a., skull stripping, is a key step in the brain image pre-processing pipelines. In rodents, this is mostly achieved by manually editing a brain mask, which is time-consuming and operator dependent. Automating this step is particularly challenging in rodents as compared to humans, because of differences in brain/scalp tissue geometry, image resolution with respect to brain-scalp distance, and tissue contrast around the skull. In this study, we proposed a deep-learning-based framework, U-Net, to automatically identify the rodent brain boundaries in MR images. The U-Net method is robust against inter-subject variability and eliminates operator dependence. To benchmark the efficiency of this method, we trained and validated our model using both in-house collected and publicly available datasets. In comparison to current state-of-the-art methods, our approach achieved superior averaged Dice similarity coefficient to ground truth T2-weighted rapid acquisition with relaxation enhancement and T2 *-weighted echo planar imaging data in both rats and mice (all p < 0.05), demonstrating robust performance of our approach across various MRI protocols.

show abstract

Section: Introductionmentioning

confidence: 99%

Automatic Skull Stripping of Rat and Mouse Brain MRI Data Using U-Net

et al. 2020

View full text Add to dashboard Cite

show abstract

“…• Random Forest-based predictive model (Bendazzoli et al, 2019): The Random Forest is an ensemble learning technique that builds on a simple decision tree's functionality by aggregating multiple decision trees' results a voting rule (Qamar et al, 2016). It has two main advantages: one, the random forest resamples the training data with replacement and reduces the variance in classification and restricts the number of features in a tree, thus providing a more optimized variable importance result.…”

Section: Proposed Model For Workflowmentioning

confidence: 99%

Deep Neural Network and Time Series Approach for Finance Systems

Srivastava

Zhang

Eachempati

2021

Journal of Organizational and End User Computing

View full text Add to dashboard Cite

The stock market is an aggregation of investor sentiment that affects daily changes in stock prices. Investor sentiment remained a mystery and challenge over time, inviting researchers to comprehend the market trends. The entry of behavioral scientists in and around the 1980s brought in the market trading's human dimensions. Shortly after that, due to the digitization of exchanges, the mix of traders changed as institutional traders started using algorithmic trading (AT) on computers. Nevertheless, the effects of investor sentiment did not disappear and continued to intrigue market researchers. Though market sentiment plays a significant role in timing investment decisions, classical finance models largely ignored the role of investor sentiment in asset pricing. For knowing if the market price is value-driven, the investor would isolate components of irrationality from the price, as reflected in the sentiment. Investor sentiment is an expression of irrational expectations of a stock's risk-return profile that is not justified by available information. In this context, the paper aims to predict the next-day trend in the index prices for the centralized Indian National Stock Exchange (NSE) deploying machine learning algorithms like support vector machine, random forest, gradient boosting, and deep neural networks. The training set is historical NSE closing price data from June 1st, 2013-June 30th, 2020. Additionally, the authors factor technical indicators like moving average (MA), moving average convergence-divergence (MACD), K (%) oscillator and corresponding three days moving average D (%), relative strength indicator (RSI) value, and the LW (R%) indicator for the same period. The predictive power of deep neural networks over other machine learning techniques is established in the paper, demonstrating the future scope of deep learning in multi-parameter time series prediction.

show abstract

“…Active contour model has been used to segment and delineate boundaries in different types of medical images such as echocardiographs [13], MRI [14], and Computed Tomography (CT) [15]. It is a popular class of image segmentation and boundary delineation method due to its ability to fit a curve to an object boundary by iteratively expanding or contracting its boundary estimate.…”

Section: Active Contour Modelmentioning

confidence: 99%

Contour evolution method for precise boundary delineation of medical images

et al. 2020

View full text Add to dashboard Cite

Image segmentation is an important precursor to boundary delineation of medical images. One of the major challenges in applying automatic image segmentation in medical images is the imperfection in the imaging process which can result in inconsistent contrast and brightness levels, and low image sharpness and vanishing boundaries. Although recent advances in deep learning produce vast improvements in the quality of image segmentation, the accuracy of segmentation around object boundaries still requires improvement. We developed a new approach to contour evolution that is more intuitive but shares some common principles with the active contour model method. The method uses two concepts, namely the boundary grid and sparse boundary representation, as an implicit and explicit representation of the boundary points. We tested our method using lumbar spine MRI images of 515 patients. The experiment results show that our method performs up to 10.2 times faster and more flexible than the geodesic active contours method. Using BF-score contour-based metric, we show that our method improves the boundary accuracy from 74% to 84% as opposed to 63% by the latter method.

show abstract

Automatic rat brain segmentation from MRI using statistical shape models and random forest

Cited by 8 publications

References 7 publications

Automatic Skull Stripping of Rat and Mouse Brain MRI Data Using U-Net

Automatic Skull Stripping of Rat and Mouse Brain MRI Data Using U-Net

Deep Neural Network and Time Series Approach for Finance Systems

Contour evolution method for precise boundary delineation of medical images

Contact Info

Product

Resources

About