UNet++: A Nested U-Net Architecture for Medical Image Segmentation
By Zongwei Zhou, Mahfuzur R. Siddiquee, Nima Tajbakhsh, Jianming Liang. 2018
9,204 citations recorded by the source corpus.
In this paper, we present UNet++, a new, more powerful architecture for medical image segmentation. Our architecture is essentially a deeply-supervised encoder-decoder network where the encoder and decoder sub-networks are connected through a series of nested, dense skip pathways. The re-designed skip pathways aim at reducing the semantic gap between the feature maps of the encoder and decoder sub-networks. We argue that the optimizer would deal with an easier learning task when the feature maps from the decoder and encoder networks are semantically similar. We have evaluated UNet++ in comparison with U-Net and wide U-Net architectures across multiple medical image segmentation tasks: nodule segmentation in the low-dose CT scans of chest, nuclei segmentation in the microscopy images, liver segmentation in abdominal CT scans, and polyp segmentation in colonoscopy videos. Our experiments demonstrate that UNet++ with deep supervision achieves an average IoU gain of 3.9 and 3.4 points over U-Net and wide U-Net, respectively.
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
- Luc Van Gool
- Ling Shao
- Ming-Hsuan Yang
- Ming-Ming Cheng
- Dacheng Tao
- Xiaogang Wang
- Alan L. Yuille
- Jianbing Shen
In Radiology, Nuclear Medicine and Imaging
- Vishal M. Patel
- Jeya Maria Jose Valanarasu
- Radu Marculescu
- Md Mostafijur Rahman
- Deng-Ping Fan
- Chenxi Duan
- Ling Shao
- Shunyi Zheng
Citations in the corpus
- Continual Barlow Twins: Continual Self-Supervised Learning for Remote Sensing Semantic Segmentation (2022, methodology): “For the semantic segmentation task, we use UNet++ [9], with the Squeeze and Excitation strategy [27] and the softmax function as activation on the last layer.”
- Deep High-Resolution Representation Learning for Visual Recognition (2019, background): “…convolutional units [64], [111], [180], as well as sending information from low-resolution skip connections to high-resolution skip connections [189] or exchanging information between them [49]; studying the details of the up-sample process [147]; combining multi-scale pyramid representations…”
- OCNet: Object Context Network for Scene Parsing (2018, neutral): “7 UNet++ (Zhou et al., 2018) ResNet-101 47.”
- Image Segmentation Using Deep Learning: A Survey (2020, methodology): “[48] developed a nested U-Net architecture, Zhang et al.”
- Medical Image Segmentation Review: The Success of U-Net (2022, background): “SCE Attention U-Net [70] UNet++ [11], [67] RA-Unet [71] BCDU-Net [74] U-Net3+ [68] BiO-Net [69] Skip connections are defined as connections in a neural network that do not connect two following layers but instead skip over at least one layer.”
- Context-aware Cross-level Fusion Network for Camouflaged Object Detection (2021, neutral): “…method with 14 state-of-the-art COD methods, including FPN [Lin et al., 2017], MaskRCNN [He et al., 2017], PSPNet [Zhao et al., 2017], UNet++ [Zhou et al., 2018], PiCANet [Liu et al., 2018], MSRCNN [Huang et al., 2019a], BASNet [Qin et al., 2019], PFANet [Zhao and Wu, 2019], CPD [Wu et al.,…”
- Unbox the black-box for the medical explainable AI via multi-modal and multi-centre data fusion: A mini-review, two showcases and beyond (2021, methodology): “We can observe that U-Net [170] could hardly predict the ventricles on thin-slice images, while U-Net++ [184] was able to segment the left lateral and right lateral ventricles by taking advantage of dense connections of the intermediate feature maps.”
- TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation (2021, background): “magnetic resonance (MR) [17], organ segmentation from computed tomography (CT) [7, 9, 18,20] and polyp segmentation [21] from colonoscopy videos.”
- U-Net Transformer: Self and Cross Attention for Medical Image Segmentation (2021, neutral): “Currently, state-of-the-art methods rely on Fully Convolutional Networks (FCNs), such as U-Net and variants [11,2,9,21].”
- SemiCDNet: A Semisupervised Convolutional Neural Network for Change Detection in High Resolution Remote-Sensing Images (2020, methodology): “To overcome the semantic gaps between encoder and decoder parts, many extensions of UNet have been proposed, such as MultiResUnet [54], MDUUNet [55], and UNet++ [49], which have opened up new perspectives of improving semantic segmentation performance for RS community.”
- Camouflaged Object Detection (2020, methodology): “GOD vs. SOD Baselines.”
- Remote Sensing Image Change Detection With Transformers (2021, methodology): “7) SNUNet [14]: Multiscale feature concatenation method, which combines the Siamese network and NestedUNet [55] to extract HR high-level features.”
- SegDiff: Image Segmentation with Diffusion Probabilistic Models (2021, methodology): “Method Dice mIoU FCN [3] 28.84 28.71 U-Net [38] 79.43 65.99 U-Net++ [53] 79.49 66.04 Res-UNet [48] 79.49 66.07 A.A U-Net [46] 76.83 62.49 MedT [45] 79.55 66.17 Ours 81.59 69.00 The performance of the mIoU segmentation metric as a function of the number of iterations is presented for the three…”
- Infrared and Visible Image Fusion via Decoupling Network (2022, background): “Li et al. [6] introduced a Nest network [28] to construct an encoder from a multiscale perspective, which preserves a large amount of information about the source images.”