Splenomegaly segmentation using global convolutional kernels and conditional generative adversarial networks

Huo, Yuankai; Xu, Zhoubing; Bao, Shunxing; Bermúdez, Camilo; Plassard, Andrew J.; Liu, Jiaqi; Yao, Yuang; Assad, Albert; Abramson, Richard G.; Landman, Bennett A.

doi:10.1117/12.2293406

Cited by 38 publications

(49 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first step in the ALARM pipeline is to obtain whole liver segmentation. The previously proposed SS‐Net deep convolutional neural network was employed to achieve whole liver segmentation (Fig. ).…”

Section: Methodsmentioning

confidence: 99%

Fully automatic liver attenuation estimation combing CNN segmentation and morphological operations

Huo¹,

Terry²,

Wang³

et al. 2019

Medical Physics

Self Cite

View full text Add to dashboard Cite

Purpose Manually tracing regions of interest (ROIs) within the liver is the de facto standard method for measuring liver attenuation on computed tomography (CT) in diagnosing nonalcoholic fatty liver disease (NAFLD). However, manual tracing is resource intensive. To address these limitations and to expand the availability of a quantitative CT measure of hepatic steatosis, we propose the automatic liver attenuation ROI‐based measurement (ALARM) method for automated liver attenuation estimation. Methods The ALARM method consists of two major stages: (a) deep convolutional neural network (DCNN)‐based liver segmentation and (b) automated ROI extraction. First, liver segmentation was achieved using our previously developed SS‐Net. Then, a single central ROI (center‐ROI) and three circles ROI (periphery‐ROI) were computed based on liver segmentation and morphological operations. The ALARM method is available as an open source Docker container (https://github.com/MASILab/ALARM). Results Two hundred and forty‐six subjects with 738 abdomen CT scans from the African American‐Diabetes Heart Study (AA‐DHS) were used for external validation (testing), independent from the training and validation cohort (100 clinically acquired CT abdominal scans). From the correlation analyses, the proposed ALARM method achieved Pearson correlations = 0.94 with manual estimation on liver attenuation estimations. When evaluating the ALARM method for detection of nonalcoholic fatty liver disease (NAFLD) using the traditional cut point of < 40 HU, the center‐ROI achieved substantial agreements (Kappa = 0.79) with manual estimation, while the periphery‐ROI method achieved “excellent” agreement (Kappa = 0.88) with manual estimation. The automated ALARM method had reduced variability compared to manual measurements as indicated by a smaller standard deviation. Conclusions We propose a fully automated liver attenuation estimation method termed ALARM by combining DCNN and morphological operations, which achieved “excellent” agreement with manual estimation for fatty liver detection. The entire pipeline is implemented as a Docker container which enables users to achieve liver attenuation estimation in five minutes per CT exam.

show abstract

Section: Methodsmentioning

confidence: 99%

Fully automatic liver attenuation estimation combing CNN segmentation and morphological operations

Huo¹,

Terry²,

Wang³

et al. 2019

Medical Physics

Self Cite

View full text Add to dashboard Cite

show abstract

“…Our previous Spleen Segmentation Network (SSNet) [7] was trained by 75 CT scans with normal spleens to see if such network was able to segment 19 splenomegaly CT scans. Then, two baseline methods were employed using 19 splenomegaly CT scans in both training and testing.…”

Section: Ct Segmentation With Ct Manual Labelsmentioning

confidence: 99%

“…Previous automated methods have been proposed to perform segmentation on normal spleens [1,2] and with splenomegaly [3][4][5]. Recently, deep convolutional neural network (DCNN) based methods have been used in splenomegaly and shown superior performance [6,7]. However, one major limitation of deploying DCNN methods is that one typically has to manually trace a new set of training data when segmenting organs in a new imaging modality or segmenting abnormal organs from a new disease cohort.…”

Section: Introductionmentioning

confidence: 99%

Adversarial synthesis learning enables segmentation without target modality ground truth

Huo

Bao

et al. 2018

2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018)

Self Cite

119

View full text Add to dashboard Cite

A lack of generalizability is one key limitation of deep learning based segmentation. Typically, one manually labels new training images when segmenting organs in different imaging modalities or segmenting abnormal organs from distinct disease cohorts. The manual efforts can be alleviated if one is able to reuse manual labels from one modality (e.g., MRI) to train a segmentation network for a new modality (e.g., CT). Previously, two stage methods have been proposed to use cycle generative adversarial networks (CycleGAN) to synthesize training images for a target modality. Then, these efforts trained a segmentation network independently using synthetic images. However, these two independent stages did not use the complementary information between synthesis and segmentation. Herein, we proposed a novel end-to-end synthesis and segmentation network (EssNet) to achieve the unpaired MRI to CT image synthesis and CT splenomegaly segmentation simultaneously without using manual labels on CT. The end-to-end EssNet achieved significantly higher median Dice similarity coefficient (0.9188) than the two stages strategy (0.8801), and even higher than canonical multi-atlas segmentation (0.9125) and ResNet method (0.9107), which used the CT manual labels.

show abstract

“…This differs from aforementioned works in the sense that the additional loss term is being learned by the discriminator rather than having fixed hand-crafted loss terms. The same mechanism was later applied to image-to-image translation [20], medical image analysis [5,6,7,8,16,21,22,23] and other segmentation tasks [24]. In contrast to our work, this formulation of adversarial training does not use the pairing information of images and labels.…”

Section: Related Workmentioning

confidence: 99%

EL-GAN: Embedding Loss Driven Generative Adversarial Networks for Lane Detection

Ghafoorian

Nugteren

Baka

et al. 2019

Lecture Notes in Computer Science

165

128

View full text Add to dashboard Cite

Convolutional neural networks have been successfully applied to semantic segmentation problems. However, there are many problems that are inherently not pixel-wise classification problems but are nevertheless frequently formulated as semantic segmentation. This ill-posed formulation consequently necessitates hand-crafted scenario-specific and computationally expensive post-processing methods to convert the per pixel probability maps to final desired outputs. Generative adversarial networks (GANs) can be used to make the semantic segmentation network output to be more realistic or better structure-preserving, decreasing the dependency on potentially complex post-processing. In this work, we propose EL-GAN: a GAN framework to mitigate the discussed problem using an embedding loss. With EL-GAN, we discriminate based on learned embeddings of both the labels and the prediction at the same time. This results in much more stable training due to having better discriminative information, benefiting from seeing both 'fake' and 'real' predictions at the same time. This substantially stabilizes the adversarial training process. We use the TuSimple lane marking challenge to demonstrate that with our proposed framework it is viable to overcome the inherent anomalies of posing it as a semantic segmentation problem. Not only is the output considerably more similar to the labels when compared to conventional methods, the subsequent post-processing is also simpler and crosses the competitive 96% accuracy threshold.

show abstract

Splenomegaly segmentation using global convolutional kernels and conditional generative adversarial networks

Cited by 38 publications

References 28 publications

Fully automatic liver attenuation estimation combing CNN segmentation and morphological operations

Fully automatic liver attenuation estimation combing CNN segmentation and morphological operations

Adversarial synthesis learning enables segmentation without target modality ground truth

EL-GAN: Embedding Loss Driven Generative Adversarial Networks for Lane Detection

Contact Info

Product

Resources

About