3D Deeply Supervised Network for Automatic Liver Segmentation from CT Volumes

By Qi-Yan Dou, Hao Chen, Yueming Jin, Lequan Yu, Jing Qin, P. Heng. 2016

370 citations recorded by the source corpus.

Automatic liver segmentation from CT volumes is a crucial prerequisite yet challenging task for computer-aided hepatic disease diagnosis and treatment. In this paper, we present a novel 3D deeply supervised network (3D DSN) to address this challenging task. The proposed 3D DSN takes advantage of a fully convolutional architecture which performs efficient end-to-end learning and inference. More importantly, we introduce a deep supervision mechanism during the learning process to combat potential optimization difficulties, and thus the model can acquire a much faster convergence rate and more powerful discrimination capability. On top of the high-quality score map produced by the 3D DSN, a conditional random field model is further employed to obtain refined segmentation results. We evaluated our framework on the public MICCAI-SLiver07 dataset. Extensive experiments demonstrated that our method achieves competitive segmentation results to state-of-the-art approaches with a much faster processing speed.

Subfields: Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Resource Optimized Neural Architecture Search for 3D Medical Image Segmentation (2019, methodology): “The architecture combines the decoder 1 × 1 × 1 convolution skip connections in DeepLabV3+[2] and the deep supervision scheme proposed in [3].”
A survey on deep learning in medical image analysis (2017, methodology): “Dou et al. (2016b) CNN Microbleed detection; 3D FCN for candidate segmentation followed by a 3D CNN as false positive reduction”
Be Your Own Teacher: Improve the Performance of Convolutional Neural Networks via Self Distillation (2019, background): “With the help of convolutional neural networks, applications such as image classification [22, 34] ,object detection [28], and semantic segmentation [7, 40] are developing at an unprecedented speed nowadays.”
Deep Learning Techniques for Medical Image Segmentation: Achievements and Challenges (2019, background): “Dou et al. in [22] applied deeply supervised approaches to segment the 3D liver CT volumes.”
Volumetric ConvNets with Mixed Residual Connections for Automated Prostate Segmentation from 3D MR Images (2017, methodology): “In addition, we also integrate a deep supervision mechanism (Lee et al. 2015; Dou et al. 2016) in our volumetric ConvNet to accelerate its convergence speed.”
CNN-based Segmentation of Medical Imaging Data (2017, methodology): “The two most common approaches are training a CNN on patches extracted from images and doing inference by sliding the CNN across all pixels of the network, predicting one pixel in each forward pass [7, 34] and training an FCN on full images or large sections of images [35, 43, 32, 4, 8, 19, 13].”
UNETR: Transformers for 3D Medical Image Segmentation (2021, methodology): “CNN-based Segmentation Networks : Since the introduction of the seminal U-Net [36], CNN-based networks have achieved state-of-the-art results on various 2D and 3D various medical image segmentation tasks [15, 54, 49, 17, 28].”
Spatial aggregation of holistically‐nested convolutional neural networks for automated pancreas localization and segmentation☆ (2017, methodology): “Alternatively, there are recent studies on directly using 3D convolutional neural networks for liver, brain segmentation [32], [33] and volumetric vascular boundary detection [34].”
H-DenseUNet: Hybrid Densely Connected UNet for Liver and Tumor Segmentation From CT Volumes (2018, background): “Moreover, compared with other 3D networks [10], [18], our method probes 3D contexts efficiently.”
RA-UNet: A Hybrid Deep Attention-Aware Network to Extract Liver and Tumor in CT Scans (2018, background): “The 3D deeply supervised network (DSN), which had a much faster convergence and better discrimination capability, could be extended to other medical applications (Dou et al., 2016).”
The Liver Tumor Segmentation Benchmark (LiTS) (2019, methodology): “good results on SLIVER07 both combining a simultaneous liver detection and probabilistic segmentation using 3D CNN with accuracy refinement of the segmentation afterwards using graph cut[57], 3D CNN in combination with random fields [58], 3D CNN following a surface”
VoxResNet: Deep voxelwise residual networks for brain segmentation from 3D MR images (2017, neutral): “…other type of methods employed real 3D CNNs to detect or segment objects from volumetric data and demonstrated compelling performance (Dou et al.; Dou et al., 2016a, 2016b; Cicek et al., 2016; Yu et al., 2017; Merkow et al., 2015; Milletari et al., 2016; Kamnitsas et al., 2017; Chen et al.,…”
UNETR++: Delving Into Efficient and Accurate 3D Medical Image Segmentation (2022, background): “In the case of 3D medical image segmentation [8,10,11,25, 31], the full volumetric image is typically processed as a sequence of 2D slices.”
CE-Net: Context Encoder Network for 2D Medical Image Segmentation (2019, background): “[47] proposed 3D deeply supervised network (3D DSN) to automatically segment lung in CT volumes.”
UNet++: Redesigning Skip Connections to Exploit Multiscale Features in Image Segmentation (2019, background): “Dou et al. [49] introduce a deep supervision by combining predictions from varying resolutions of feature maps, suggesting that it can combat potential optimization difﬁculties and thus reach faster convergence rate and more powerful discrimination capability.”