Fast Fully Automatic Segmentation of the Human Placenta from Motion Corrupted MRI
read more
Citations
A survey on deep learning in medical image analysis
DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentation
Ensembles of Multiple Models and Architectures for Robust Brain Tumour Segmentation
Ensembles of Multiple Models and Architectures for Robust Brain Tumour Segmentation
A deep learning framework for supporting the classification of breast lesions in ultrasound images.
References
SLIC Superpixels Compared to State-of-the-Art Superpixel Methods
Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials
Efficient Multi-Scale 3D CNN with Fully Connected CRF for Accurate Brain Lesion Segmentation
CPR - curved planar reformation
Mean Curvature Skeletons
Related Papers (5)
Frequently Asked Questions (21)
Q2. What are the future works in "Fast fully automatic segmentation of the human placenta from motion corrupted mri" ?
Moreover, the authors extend their framework scope to real clinical applications by compensating motion artifacts using slice to volume registration techniques, as well as providing a novel standardized view into the placental structures using skeleton extraction and curved planar reformation. In future work the authors will investigate the potential use of the standardized placenta views for image-based classification and automatic detection of abnormalities.
Q3. What are the quantitative measurements required for identifying abnormalities?
quantitative measurements such as placental volume and surface attachment to the uterine wall, are required for identifying abnormalities.
Q4. How many test subjects are used in this experiment?
The network is evaluated with 2-fold cross validation, 10 test subjects from Dataset II, and 44+10 training subjects from Dataset The authorand Dataset II.
Q5. What is the proposed framework for the placenta from motion-corrupted ?
The proposed framework adopts convolutional neural networks (CNNs) as a strong classifier for image segmentation followed by a conditional random field (CRF) for refinement.
Q6. What is the role of the placenta in the fetal development?
Recent work [8] has shown that magnetic resonance imaging (MRI) can be used for the evaluation of the placenta during both normal and high-risk pregnancies.
Q7. What is the way to measure the placenta from a motion-free ?
Even though ultrasound is fast enough to acquire a motion free volume, the lack of structural information and weak tissue gradients make it only useful for volume measurements.
Q8. How do the authors deal with the variations of the placenta’s appearance?
In order to deal with the variations of the placenta’s appearance, the authors apply data augmentation for training by flipping the image around the main 3D axes (maternal orientation).
Q9. What is the CRF model used to define the pairwise edge potentials?
The authors use a 3D fully connected CRF model [7, 4] which applies a linear combination of Gaussian kernels to define the pairwise edge potentials.
Q10. What is the advantage of the multi-scale architecture?
This multi-scale architecture has the advantage of capturing larger 3D contextual information, which is essential for detecting highly variable organs.
Q11. What are the main reasons for the difficulty in acquiring 3D images?
3D data acquisition and subsequent automatic segmentation is challenging because maternal respiratory motion and fetal movements displace the overall anatomy, which causes motion artifacts between individual slices as shown in Fig.
Q12. What is the proposed approach for measuring the placenta from motion free MRI?
The proposed approach combines a 3D multi-scale CNN architecture for segmentation with a 3D dense CRF for segmentation refinement.
Q13. What is the way to tackle the motion artifacts caused by the placent?
To tackle these motion artifacts caused by fetal and maternal movements the authors combine their segmentation framework with flexible motion compensation algorithm based on patch-to-volume registration (PVR) [3].
Q14. How many fetuses are in the dataset II?
Dataset II contains 22 MR scans of healthy fetuses and fetuses with intrauterine fetal growth restriction (IUGR) at gestational age between 20–38 weeks.
Q15. What is the recent work in fetal MRI?
Most previous work in fetal MRI was focused on brain segmentation [2] and very recently has been extended to localize other fetal organs [6].
Q16. What is the morphological structure of the placenta?
It is supported by a mean-curvature flow skeleton [10] generated from the triangulated polygonal mesh of the placenta segmentation and textured similar to curved planar reformation [5], see Figure 3.
Q17. What is the main reason why the authors use a multi-scale architecture?
Despite the fact that the multi-scale architecture can interpret contextual information, inference is subject to misclassification and errors.
Q18. How many subjects are tested in the first experiment?
In a second experiment [exp-2], the authors train the CNN using the whole 44 subject from Dataset The authorand test it on the 22 subjects from Dataset II.
Q19. What is the placental volume measurement method?
The authors also show howthe resulting reconstructed volume can be used to provide a novel standardized view into the placental structures by applying shape skeleton extraction and curved planar reformation for shape abstraction.
Q20. What is the way to segment the placenta?
Although this tissue classification approach is capable of segmenting the placenta robustly, the segmentation is still subject to inter-slice motion artifacts.
Q21. What is the purpose of this paper?
In this paper the authors propose for the first time a fully automatic segmentation framework for the placenta from motion corrupted fetal MRI.