Automatic 3D model-based method for liver segmentation in MRI based on active contours and total variation minimization

Bereciartua, Arantza; Picón, Artzai; Galdrán, Adrián; Iriondo, Pedro

doi:10.1016/j.bspc.2015.04.005

Cited by 25 publications

(19 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, several approaches for automatic liver segmentation in MRI have been proposed. A 3D liver model guided by a precomputed probability map was used by Bereciartua et al [4], which achieved a mean Dice of 0.90 for patients with healthy livers. Le et al used a histogram-based liver segmentation with a subsequent geodesic active contour refinement step and reported a mean Dice of 0.91 [5].…”

Section: Introductionmentioning

confidence: 99%

Reducing inter-observer variability and interaction time of MR liver volumetry by combining automatic CNN-based liver segmentation and manual corrections

et al. 2019

View full text Add to dashboard Cite

Purpose To compare manual corrections of liver masks produced by a fully automatic segmentation method based on convolutional neural networks (CNN) with manual routine segmentations in MR images in terms of inter-observer variability and interaction time. Methods For testing, patient’s precise reference segmentations that fulfill the quality requirements for liver surgery were manually created. One radiologist and two radiology residents were asked to provide manual routine segmentations. We used our automatic segmentation method Liver-Net to produce liver masks for the test cases and asked a radiologist assistant and one further resident to correct the automatic results. All observers were asked to measure their interaction time. Both manual routine and corrected segmentations were compared with the reference annotations. Results The manual routine segmentations achieved a mean Dice index of 0.95 and a mean relative error (RVE) of 4.7%. The quality of liver masks produced by the Liver-Net was on average 0.95 Dice and 4.5% RVE. Liver masks resulting from manual corrections of automatically generated segmentations compared to routine results led to a significantly lower inter-observer variability (mean per case absolute RVE difference across observers 0.69%) when compared to manual routine ones (2.75%). The mean interaction time was 2 min for manual corrections and 10 min for manual routine segmentations. Conclusions The quality of automatic liver segmentations is on par with those from manual routines. Using automatic liver masks in the clinical workflow could lead to a reduction of segmentation time and a more consistent liver volume estimation across different observers.

show abstract

Section: Introductionmentioning

confidence: 99%

Reducing inter-observer variability and interaction time of MR liver volumetry by combining automatic CNN-based liver segmentation and manual corrections

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Furthermore, none of the authors of previous works have made their own datasets public. Since the developed algorithm is meant to be capable of segmenting of MRI images available in clinics, we collected datasets from several medical centers and hospitals; where the MRI images are all of type T1; a type preferred by radiologists for its quality [4]. All abdominal MRI datasets have been acquired of the liver.…”

Section: Resultsmentioning

confidence: 99%

“…1(d). Hence, most of the time, clearly defined edges are not visible on many sides of the liver [4]- [6].…”

Section: Introductionmentioning

confidence: 99%

An Adaptive Method for Fully Automatic Liver Segmentation in Medical MRI-Images

Mohamed

Seada

Hamdy

et al. 2017

IJCA

View full text Add to dashboard Cite

Despite the importance of the liver segmentation in the medical images for efficient noninvasive diagnosis, few studies found in the literatures for fully automated methods for liver segmentation in Magnetic Resonance Imaging (MRI) compared to that in Computed Tomography (CT) scans. Motivated by this, we propose an adaptive fully automatic liver segmentation method for MRI images based on thresholding and Bayesian classification. Bayesian classifications have proved to be highly robust to various image degradations. It only requires a small amount of training data to estimate the parameters necessary for classification, which is a huge advantage in medical applications. Furthermore, the Bayesian model is robust when large uncertainties are involved in medical image analysis problems. The proposed method is successfully tested on many MRI cases acquired from different patients, in various sizes. Experiments proved the robustness of the proposed liver automatic segmentation process even on data from different scanner types. The segmentation accuracy of the model has a mean Dice Similarity Coefficient (DSC) of 95.5% for MRI datasets.

show abstract

“…Prior to the use of deep learning (DL), segmentation would most frequently be based multi-atlas approaches. [1] uses 3D models of the liver and probability maps, [2] is based on histograms to segment the liver, followed by active contours for refinement, [3] applies watershed together with active contours. Then deep learning-based segmentation revolutionized the field.…”

Section: Related Workmentioning

confidence: 99%

Investigation of Loss Functions for Improving Deep Segmentation of Abdominal Organs from MRI

Furtado¹

2020

Preprint

View full text Add to dashboard Cite

Segmentation of Magnetic Resonance Images (MRI) of abdominal organs is useful for analysis prior to surgical procedures and for further processing. Deep Learning (DL) has become the standard, researchers have proposed improvements that include multiple views, ensembles and voting. Loss function alternatives, while being crucial to guide automated learning, have not been compared in detail. In this work we analyze limitations of popular metrics and their use as loss, study alternative loss variations based on those and other modifications and search for the best approach. An experimental setup was necessary to assess the alternatives. Results for the top scoring network and top scoring loss show improvements between 2 and 11 percentage points (pp) in Jaccard Index (JI), depending on organ and patient (sequence), for a total of 22 pp over 4 organs, all this being obtained just by choosing the best performing loss function instead of cross-entropy or dice. Our results apply directly to MRI of abdominal organs, with important practical implications for other architectures, as they can be applied easily to any of them. They also show the worth of variants of loss function and loss tuning, with future work needed to generalize and test in other contexts.

show abstract

Automatic 3D model-based method for liver segmentation in MRI based on active contours and total variation minimization

Cited by 25 publications

References 31 publications

Reducing inter-observer variability and interaction time of MR liver volumetry by combining automatic CNN-based liver segmentation and manual corrections

Reducing inter-observer variability and interaction time of MR liver volumetry by combining automatic CNN-based liver segmentation and manual corrections

An Adaptive Method for Fully Automatic Liver Segmentation in Medical MRI-Images

Investigation of Loss Functions for Improving Deep Segmentation of Abdominal Organs from MRI

Contact Info

Product

Resources

About