Papers › U²-Net: Going Deeper with Nested U-Structure for Salient Object Detection

U²-Net: Going Deeper with Nested U-Structure for Salient Object Detection

18 May 2020arXiv:2005.09007archive 2025-07-28

Xuebin Qin, Zichen Zhang, Chenyang Huang, Masood Dehghan, Osmar R. Zaiane, Martin Jagersand

In this paper, we design a simple yet powerful deep network architecture, U²-Net, for salient object detection (SOD). The architecture of our U²-Net is a two-level nested U-structure. The design has the following advantages: (1) it is able to capture more contextual information from different scales thanks to the mixture of receptive fields of different sizes in our proposed ReSidual U-blocks (RSU), (2) it increases the depth of the whole architecture without significantly increasing the computational cost because of the pooling operations used in these RSU blocks. This architecture enables us to train a deep network from scratch without using backbones from image classification tasks. We instantiate two models of the proposed architecture, U²-Net (176.3 MB, 30 FPS on GTX 1080Ti GPU) and U²-Net^† (4.7 MB, 40 FPS), to facilitate the usage in different environments. Both models achieve competitive performance on six SOD datasets. The code is available: https://github.com/NathanUA/U-2-Net.

PaperPDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2005.09007")

Code

Syntology Ran 4 of 42 code samples harvested from 11 repositories linked to this paper; 38 have no recorded run. Of those that ran: 1 ran · honoured contract; 1 ran · violated contract; 2 ran · fixture could not drive it.

By repository: official repository: 2 samples from 1 repository, 0 ran; community (archive-listed): 40 samples from 10 repositories, 4 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

29 repositories listed; official and paper-mentioned ones first.

NathanUA/U-2-Net officialmentioned in papermentioned on GitHubpytorchApache-2.0 report
He-jerry/U2Net-Tensorflow mentioned on GitHubtf report
PPPrior/u2net mentioned on GitHubpytorch report
RoadoneP/aloha mentioned on GitHubpytorchMIT report
alililia/u2net_Ascend mentioned on GitHubmindsporeApache-2.0 report
alililia/u2net_GPU mentioned on GitHubmindsporeApache-2.0 report
chjort/u2net mentioned on GitHubtf report
dengpingfan/fsgan mentioned on GitHubpytorch report
hahahappyboy/GANForCartoon mentioned on GitHubpytorchGPL-2.0 report
hhaAndroid/u2net mentioned on GitHubpytorch report
kingcong/u2net mentioned on GitHubmindsporenot reachable when probed 2026-09-17 — repositories for recent papers often appear after camera-ready report
levindabhi/cloth-segmentation mentioned on GitHubpytorchMIT report
lucidrains/imagen-pytorch mentioned on GitHubpytorchMIT report
nadermx/backgroundremover mentioned on GitHubpytorchMIT report
patharanordev/background-remover mentioned on GitHubpytorchMIT report
qooba/aiscissors mentioned on GitHubMIT report
schenock/u2net-generic mentioned on GitHubpytorchGPL-3.0 report
singhaman092/rembg mentioned on GitHubpytorchMIT report
sk1123344/U-2-NET-and-U-NET mentioned on GitHubpytorch report
soxHenry433/TF2_U2Model mentioned on GitHubtf report
vaibhavmit074/remove-bg-2 mentioned on GitHubpytorchApache-2.0 report
xuebinqin/U-2-Net mentioned on GitHubpytorch report
yangyucheng000/u2net mentioned on GitHubmindspore report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

42 samples harvested; 4 ran; 1 honoured the contract we drafted; 38 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

1ran · honoured contract
1ran · violated contract
2ran · fixture could not drive it
38unverified

Licence: 2 of the 42 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 11 repositories linked to this paper, official or community; each sample names its own and says which. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

crop_face NathanUA/U-2-Net/u2net_portrait_demo.py official repository unverified Apache-2.0 (permissive) · b42507b856af3374 · report
detect_single_face NathanUA/U-2-Net/u2net_portrait_demo.py official repository unverified Apache-2.0 (permissive) · e5a6c2c113b235f3 · report
bce_loss_all sk1123344/U-2-NET-and-U-NET/u2net_train_test.py community (archive-listed) ran · fixture could not drive it no licence file found · pointer only · 81a853f579032ae4 · report
muti_bce_loss_fusion xuebinqin/U-2-Net/u2net_train.py community (archive-listed) ran · fixture could not drive it Apache-2.0 (permissive) · fc74645243d46d51 · report
normPRED hhaAndroid/u2net/u2net_test.py community (archive-listed) ran · honoured contract fingerprinted Apache-2.0 (permissive) · 89e4412c100b0f9b · report
str2bool RoadoneP/aloha/eval_image.py community (archive-listed) ran · violated contract MIT (permissive) · f017532fc389cbfe · report
REBNCONV Akhilesh64/Image-Segmentation-U-2-Net/U2-net.py community (archive-listed) unverified MIT (permissive) · 1109359fbff5405b · report
RSU yingkaisha/keras-unet-collection/keras_unet_collection/_model_u2net_2d.py community (archive-listed) unverified MIT (permissive) · c0d03d6bccec6ee1 · report
RSU4F yingkaisha/keras-unet-collection/keras_unet_collection/_model_u2net_2d.py community (archive-listed) unverified MIT (permissive) · 2e03f0a5f07cb196 · report
RSU7 Akhilesh64/Image-Segmentation-U-2-Net/U2-net.py community (archive-listed) unverified MIT (permissive) · a28583cb5a89e1f2 · report
center_size RoadoneP/aloha/layers/box_utils.py community (archive-listed) unverified MIT (permissive) · 0118ac3b2ad5818b · report
check_bidi singhaman092/rembg/other/.serverless/requirements/idna/core.py community (archive-listed) unverified MIT (permissive) · 354103a1d87fedfd · report
construct_backbone RoadoneP/aloha/backbone.py community (archive-listed) unverified MIT (permissive) · 87523070b15ff05c · report
darknetconvlayer RoadoneP/aloha/backbone.py community (archive-listed) unverified MIT (permissive) · 901e174e8a74774c · report
download_files_from_github nadermx/backgroundremover/backgroundremover/github.py community (archive-listed) unverified MIT (permissive) · b752b6ca8885fd9e · report
getData Akhilesh64/Image-Segmentation-U-2-Net/dataLoader.py community (archive-listed) unverified MIT (permissive) · 21c24045e17c6985 · report
getData Akhilesh64/Image-Segmentation-U-2-Net/Scripts_DUTS/dataLoader.py community (archive-listed) unverified MIT (permissive) · 8144398e708b679f · report
get_model patharanordev/background-remover/src/backgroundremover/bg.py community (archive-listed) unverified MIT (permissive) · 8e82437510a87611 · report
get_request singhaman092/rembg/other/utils/utils.py community (archive-listed) unverified MIT (permissive) · 71858fc2cedc576f · report
gram RoadoneP/aloha/style_utils.py community (archive-listed) unverified MIT (permissive) · 93ebd13b9e693ca3 · report
intersect RoadoneP/aloha/layers/box_utils.py community (archive-listed) unverified MIT (permissive) · 2b050648fe0422de · report
itot RoadoneP/aloha/style_utils.py community (archive-listed) unverified MIT (permissive) · e38d55db4e9ff568 · report
load_checkpoint levindabhi/cloth-segmentation/utils/saving_utils.py community (archive-listed) unverified MIT (permissive) · ae7eca3db4489538 · report
load_checkpoint_mgpu levindabhi/cloth-segmentation/utils/saving_utils.py community (archive-listed) unverified MIT (permissive) · ec0fd63b77ab92d6 · report
load_image RoadoneP/aloha/style_utils.py community (archive-listed) unverified MIT (permissive) · 4969a06ac02c02d2 · report
load_model singhaman092/rembg/src/rembg/u2net/detect.py community (archive-listed) unverified MIT (permissive) · 82b06ae41a637f16 · report
loss Akhilesh64/Image-Segmentation-U-2-Net/U2-net.py community (archive-listed) unverified MIT (permissive) · 14100923eb6556ac · report
max_workers nadermx/backgroundremover/backgroundremover/bg.py community (archive-listed) unverified MIT (permissive) · f4893b2a0f0fc410 · report
muti_bce_loss_fusion hhaAndroid/u2net/u2net_train.py community (archive-listed) unverified Apache-2.0 (permissive) · 9093b315696852d6 · report
naive_cutout nadermx/backgroundremover/backgroundremover/bg.py community (archive-listed) unverified MIT (permissive) · f250811a24b78ae9 · report
naive_cutout patharanordev/background-remover/src/backgroundremover/bg.py community (archive-listed) unverified MIT (permissive) · e3da1e070910201f · report
norm_pred nadermx/backgroundremover/backgroundremover/u2net/detect.py community (archive-listed) unverified MIT (permissive) · cfded26c9f3e75e8 · report
point_form RoadoneP/aloha/layers/box_utils.py community (archive-listed) unverified MIT (permissive) · 27dbfe70566c2e97 · report
post_request singhaman092/rembg/other/utils/utils.py community (archive-listed) unverified MIT (permissive) · b13016e1f4154113 · report
pred_norm sk1123344/U-2-NET-and-U-NET/u2net_train_test.py community (archive-listed) unverified no licence file found · pointer only · 1300ad8439b1c5bf · report
preprocess nadermx/backgroundremover/backgroundremover/u2net/detect.py community (archive-listed) unverified MIT (permissive) · c20a71a3b663ddb2 · report
preprocess singhaman092/rembg/src/rembg/u2net/detect.py community (archive-listed) unverified MIT (permissive) · c576953d98bccacc · report
tensor_for_board levindabhi/cloth-segmentation/utils/tensorboard_utils.py community (archive-listed) unverified MIT (permissive) · 30ce730ccf4be37e · report
tensor_list_for_board levindabhi/cloth-segmentation/utils/tensorboard_utils.py community (archive-listed) unverified MIT (permissive) · ae4ce56419ea054d · report
u2net_2d_base yingkaisha/keras-unet-collection/keras_unet_collection/_model_u2net_2d.py community (archive-listed) unverified MIT (permissive) · b6c716b95682bb86 · report
valid_label_length singhaman092/rembg/other/.serverless/requirements/idna/core.py community (archive-listed) unverified MIT (permissive) · 82202184fa5b7ac9 · report
valid_string_length singhaman092/rembg/other/.serverless/requirements/idna/core.py community (archive-listed) unverified MIT (permissive) · c2eaff651c95d1f4 · report

Tasks

Dichotomous Image SegmentationImage ClassificationObject DetectionRGB Salient Object DetectionSaliency DetectionSalient Object Detectionimage-classification

1 archive task tag without a task page not shown.

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Dichotomous Image Segmentation DIS-TE1 U2Net E-measure 0.801 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 U2Net HCE 224 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 U2Net MAE 0.083 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 U2Net S-Measure 0.760 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 U2Net max F-Measure 0.694 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 U2Net weighted F-measure 0.601 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 U2Net E-measure 0.833 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 U2Net HCE 490 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 U2Net MAE 0.085 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 U2Net S-Measure 0.788 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 U2Net max F-Measure 0.756 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 U2Net weighted F-measure 0.668 #8 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 U2Net E-measure 0.858 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 U2Net HCE 965 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 U2Net MAE 0.079 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 U2Net S-Measure 0.809 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 U2Net max F-Measure 0.798 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 U2Net weighted F-measure 0.707 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 U2Net E-measure 0.847 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 U2Net HCE 3653 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 U2Net MAE 0.087 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 U2Net S-Measure 0.807 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 U2Net max F-Measure 0.795 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 U2Net weighted F-measure 0.705 #7 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD U2Net E-measure 0.823 #9 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD U2Net HCE 1413 #9 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD U2Net MAE 0.090 #9 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD U2Net S-Measure 0.781 #9 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD U2Net max F-Measure 0.748 #9 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD U2Net weighted F-measure 0.656 #9 of 24 Archive leaderboard report
Saliency Detection DUT-OMRON U2-Net MAE 0.054 #3 of 5 Archive leaderboard report
Saliency Detection DUT-OMRON U2-Net+ Fwβ 0.731 #4 of 5 Archive leaderboard report
Saliency Detection DUT-OMRON U2-Net+ MAE 0.06 #4 of 5 Archive leaderboard report
Saliency Detection DUT-OMRON U2-Net+ Sm 0.837 #4 of 5 Archive leaderboard report
Saliency Detection DUT-OMRON U2-Net+ relaxFbβ 0.676 #4 of 5 Archive leaderboard report
Saliency Detection DUT-OMRON U2-Net+ {max}Fβ 0.813 #4 of 5 Archive leaderboard report
Saliency Detection HKU-IS U2-Net+ Fwβ 0.867 #3 of 3 Archive leaderboard report
Saliency Detection HKU-IS U2-Net+ MAE 0.037 #3 of 3 Archive leaderboard report
Saliency Detection HKU-IS U2-Net+ Sm 0.908 #3 of 3 Archive leaderboard report
Saliency Detection HKU-IS U2-Net+ relaxFbβ 0.794 #3 of 3 Archive leaderboard report
Saliency Detection HKU-IS U2-Net+ {max}Fβ 0.928 #3 of 3 Archive leaderboard report
Salient Object Detection ECSSD F3Net MAE 0.041 #10 of 10 Archive leaderboard report
Salient Object Detection ECSSD F3Net S-measure 0.918 #10 of 10 Archive leaderboard report
Salient Object Detection ECSSD F3Net max_F1 0.885 #10 of 10 Archive leaderboard report
Salient Object Detection HKU-IS U2Net MAE 0.031 #9 of 9 Archive leaderboard report
Salient Object Detection PASCAL-S F3Net MAE 0.086 #10 of 10 Archive leaderboard report
Salient Object Detection PASCAL-S F3Net S-measure 0.831 #10 of 10 Archive leaderboard report
Salient Object Detection PASCAL-S F3Net max_F1 0.768 #10 of 10 Archive leaderboard report
Salient Object Detection SOD U2-Net+ Fwβ 0.697 #1 of 1 Archive leaderboard report
Salient Object Detection SOD U2-Net+ MAE 0.124 #1 of 1 Archive leaderboard report
Salient Object Detection SOD U2-Net+ Sm 0.759 #1 of 1 Archive leaderboard report
Salient Object Detection SOD U2-Net+ relaxFbβ 0.559 #1 of 1 Archive leaderboard report
Salient Object Detection SOD U2-Net+ {max}Fβ 0.841 #1 of 1 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Introduced by this paper: U2-Net

U2-Net

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections