Discriminative dictionary learning for abdominal multi-organ segmentation

Tong, Tong; Wolz, Robin; Wang, Zehan; Gao, Qinquan; Misawa, Kazunari; Fujiwara, Michitaka; Mori, Kensaku; Hajnal, Joseph V.; Rueckert, Daniel

doi:10.1016/j.media.2015.04.015

Cited by 145 publications

(111 citation statements)

References 27 publications

Supporting

Mentioning

109

Contrasting

Order By: Relevance

“…We do this as MALP is an application that inherently requires establishing correspondences between images in order to propagate labels. Most state-of-the-art methods in MALP such as in [13,11] use affine registration as a first step to give an initial set of dense correspondences between the atlases and the target image before proceeding with a more sophisticated label propagation scheme. Although affine registration is less accurate than doing non-rigid registration, it is used because it is more efficient.…”

Section: Experiments and Resultsmentioning

confidence: 99%

“…Qualitative evaluation of the estimated correspondences in a simple multi-atlas propagation setting demonstrate the potential of using SVFs for estimating correspondences. We do not apply any further post-processing to improve the segmentation, such as graph-cuts, which is what is typically done in some state-of-the-art methods for segmenting abdominal datasets [13,11]. Random forests are extremely efficient during test time making them an attractive option to use for estimating correspondences in large datasets.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Supervoxel Classification Forests for Estimating Pairwise Image Correspondences

Kanavati

Tong

Misawa

et al. 2015

Machine Learning in Medical Imaging

Self Cite

View full text Add to dashboard Cite

Abstract. This paper proposes a general method for establishing pairwise correspondences, which is a fundamental problem in image analysis. The method consists of over-segmenting a pair of images into supervoxels. A forest classifier is then trained on one of the images, the source, by using supervoxel indices as voxelwise class labels. Applying the forest on the other image, the target, yields a supervoxel labelling which is then regularized using majority voting within the boundaries of the target's supervoxels. This yields semi-dense correspondences in a fully automatic, efficient and robust manner. The advantage of our approach is that no prior information or manual annotations are required, making it suitable as a general initialisation component for various medical imaging tasks that require coarse correspondences, such as, atlas/patch-based segmentation, registration, and atlas construction. Our approach is evaluated on a set of 150 abdominal CT images. In this dataset we use manual organ segmentations for quantitative evaluation. In particular, the quality of the correspondences is determined in a label propagation setting. Comparison to other state-of-the-art methods demonstrate the potential of supervoxel classification forests for estimating image correspondences.

show abstract

Section: Experiments and Resultsmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Supervoxel Classification Forests for Estimating Pairwise Image Correspondences

Kanavati

Tong

Misawa

et al. 2015

Machine Learning in Medical Imaging

Self Cite

View full text Add to dashboard Cite

show abstract

“…The probabilistic output from the forest was then used as the unary cost for graph-cut to regularise the output. Apart from poor results for the pancreas, which is difficult to segment due to its extremely deformable nature, the liver, kidney and spleen obtain good segmentation results, excluding a few outlier cases (a fully-supervised method [15] applied on the same dataset obtains a dice overlap of 94.9%, 93.6%, and 92.5% for liver, kidneys, and spleen, respectively.). Our method could potentially be used to provide coarse segmentation or mine a large dataset of medical images, with extremely minimal user input.…”

Section: Discussionmentioning

confidence: 99%

Joint Supervoxel Classification Forest for Weakly-Supervised Organ Segmentation

Kanavati

Misawa

Fujiwara

et al. 2017

Machine Learning in Medical Imaging

Self Cite

View full text Add to dashboard Cite

Abstract. This article presents an efficient method for weakly-supervised organ segmentation. It consists in over-segmenting the images into objectlike supervoxels. A single joint forest classifier is then trained on all the images, where (a) the supervoxel indices are used as labels for the voxels, (b) a joint node optimisation is done using training samples from all the images, and (c) in each leaf node, a distinct posterior distribution is stored per image. The result is a forest with a shared structure that efficiently encodes all the images in the dataset. The forest can be applied once on a given source image to obtain supervoxel label predictions for its voxels from all the other target images in the dataset by simply looking up the target's distribution in the leaf nodes. The output is then regularised using majority voting within the boundaries of the source's supervoxels. This yields sparse correspondences on an over-segmentation-based level in an unsupervised, efficient, and robust manner. Weak annotations can then be propagated to other images, extending the labelled set and allowing an organ label classification forest to be trained. We demonstrate the effectiveness of our approach on a dataset of 150 abdominal CT images where, starting from a small set of 10 images with scribbles, we perform weakly-supervised image segmentation of the kidneys, liver and spleen. Promising results are obtained.

show abstract

“…This has led to the development of numerous patch-based segmentation approaches, including several from our own group. These methods currently provide state-of-theart performance in many applications including brain ], heart [Bai et al (2013)] and abdominal organ segmentation Tong et al (2015)]. …”

Section: Semantic Imagingmentioning

confidence: 99%

Learning clinically useful information from images: Past, present and future

Rueckert

Glocker

Kainz

2016

Medical Image Analysis

Self Cite

View full text Add to dashboard Cite

Over the last decade, research in medical imaging has made significant progress in addressing challenging tasks such as image registration and image segmentation. In particular, the use of model-based approaches has been key in numerous, successful advances in methodology. The advantage of modelbased approaches is that they allow the incorporation of prior knowledge acting as a regularisation that favours plausible solutions over implausible ones. More recently, medical imaging has moved away from hand-crafted, and often explicitly designed models towards data-driven, implicit models that are constructed using machine learning techniques. This has led to major improvements in all stages of the medical imaging pipeline, from acquisition and reconstruction to analysis and interpretation. As more and more imaging data is becoming available, e.g., from large population studies, this trend is likely to continue and accelerate. At the same time new developments in machine learning, e.g., deep learning, as well as significant improvements in computing power, e.g., parallelisation on graphics hardware, offer new potential for data-driven, semantic and intelligent medical imaging. This article outlines the work of the BioMedIA group in this area and highlights some of the challenges and opportunities for future work.

show abstract

Discriminative dictionary learning for abdominal multi-organ segmentation

Cited by 145 publications

References 27 publications

Supervoxel Classification Forests for Estimating Pairwise Image Correspondences

Supervoxel Classification Forests for Estimating Pairwise Image Correspondences

Joint Supervoxel Classification Forest for Weakly-Supervised Organ Segmentation

Learning clinically useful information from images: Past, present and future

Contact Info

Product

Resources

About