Papers › Rethinking Atrous Convolution for Semantic Image Segmentation

Rethinking Atrous Convolution for Semantic Image Segmentation

17 Jun 2017arXiv:1706.05587archive 2025-07-28

Liang-Chieh Chen, George Papandreou, Florian Schroff, Hartwig Adam

In this work, we revisit atrous convolution, a powerful tool to explicitly adjust filter's field-of-view as well as control the resolution of feature responses computed by Deep Convolutional Neural Networks, in the application of semantic image segmentation. To handle the problem of segmenting objects at multiple scales, we design modules which employ atrous convolution in cascade or in parallel to capture multi-scale context by adopting multiple atrous rates. Furthermore, we propose to augment our previously proposed Atrous Spatial Pyramid Pooling module, which probes convolutional features at multiple scales, with image-level features encoding global context and further boost performance. We also elaborate on implementation details and share our experience on training our system. The proposed `DeepLabv3' system significantly improves over our previous DeepLab versions without DenseCRF post-processing and attains comparable performance with other state-of-art models on the PASCAL VOC 2012 semantic image segmentation benchmark.

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="1706.05587")

Code

Syntology Ran 3 of 7 code samples harvested from 4 repositories linked to this paper; 4 have no recorded run. Of those that ran: 1 ran · honoured contract; 1 ran · our draft was wrong; 1 ran with no contract checked.

By repository: community (archive-listed): 7 samples from 4 repositories, 3 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

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

2023-MindSpore-1/ms-code-167 mentioned on GitHubmindspore report
AutomatedAI/deeplab_inference mentioned on GitHubtf report
BebDong/MXNetSeg mentioned on GitHubmxnetApache-2.0 report
DLWK/EANet mentioned on GitHubpytorch report
IRVLab/SUIM mentioned on GitHubNOASSERTION report
IRVLab/SUIM-Net mentioned on GitHubNOASSERTION report
JWuzyk/CudaVisionProject mentioned on GitHubpytorch report
KPMG-wiseuniv/AI mentioned on GitHubpytorchGPL-3.0 report
Popcorn-sugar/Deep_v2 mentioned on GitHubtf report
Robinatp/Deeplab_Tensorflow mentioned on GitHubtf report
Syarujianai/deeplab-commented mentioned on GitHubtf report
TheTrveAnthony/no-Green mentioned on GitHubpytorch report
VainF/DeepLabV3Plus-Pytorch mentioned on GitHubpytorch report
chenmengyang/rename_later mentioned on GitHub report
chenxi116/DeepLabv3.pytorch mentioned on GitHubpytorch report
czarmanu/sentinel_lakeice mentioned on GitHubtf report
dajes/DensePose-TorchScript mentioned on GitHubpytorch report
ensta-u2is/deeplabv3plus-muad-pytorch mentioned on GitHubpytorch report
fregu856/deeplabv3 mentioned on GitHubpytorch report
giannifranchi/deeplabv3-superpixelmix mentioned on GitHubpytorch report
giovanniguidi/deeplabV3-PyTorch mentioned on GitHubpytorch report
giovanniguidi/deeplabV3_Pytorch mentioned on GitHubpytorch report
heidongxianhau/deeplab2 mentioned on GitHubtf report
it6aidl/outdoorsegmentation mentioned on GitHubpytorch report
leimao/DeepLab-V3 mentioned on GitHubtf report
leimao/DeepLab_v3 mentioned on GitHubtf report
leonardoaraujosantos/seg_atrous mentioned on GitHubpytorch report
lewandofskee/MobileMamba mentioned on GitHubpytorch report
mathildor/DeepLab-v3 mentioned on GitHubtf report
msminhas93/deeplabv3finetuning mentioned on GitHubpytorch report
naver-ai/BlendNeRF mentioned on GitHubpytorch report
osmr/imgclsmob mentioned on GitHubmxnetMIT report
rishizek/tensorflow-deeplab-v3 mentioned on GitHubtf report
samson6460/tf2_Segmentation mentioned on GitHubtf report
sharifelguindi/DeepLab mentioned on GitHubtf report
sthalles/deeplab_v3 mentioned on GitHubtf report
stigma0617/VoVNet-DeepLabV3 mentioned on GitHubpytorch report
tensorflow/models mentioned on GitHubtf report
tensorflow/models mentioned on GitHubtf report
xahidbuffon/SUIM mentioned on GitHub report
xahidbuffon/SVAM-Net mentioned on GitHubtf report
zhangzjn/emo mentioned on GitHubpytorch report
zhangzjn/emov2 mentioned on GitHubpytorch report
zxleong/GPRNet mentioned on GitHub 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

7 samples harvested; 3 ran; 1 honoured the contract we drafted; 4 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 · our draft was wrong
1ran
4unverified

Licence: 3 of the 7 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 4 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.

label_to_color_image Robinatp/Deeplab_Tensorflow/DeepLab_inference_Demo.py community (archive-listed) ran · honoured contract no licence file found · pointer only · 4634c9f4f6898b4d · report
normalize_fun samson6460/tf2_Segmentation/utils/data_processing.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · 1db5c78d80803361 · report
read_img samson6460/tf2_Segmentation/utils/data_processing.py community (archive-listed) ran · metamorphic tier: well formed no licence file found · pointer only · b5666228355ff947 · report
convert_to_separable_conv giannifranchi/deeplabv3-superpixelmix/network/_deeplab.py community (archive-listed) unverified MIT (permissive) · ce942a958c0ae909 · report
deeplabv3_hrnetv2_32 VainF/DeepLabV3Plus-Pytorch/network/modeling.py community (archive-listed) unverified MIT (permissive) · e980799df5aec203 · report
deeplabv3_hrnetv2_48 VainF/DeepLabV3Plus-Pytorch/network/modeling.py community (archive-listed) unverified MIT (permissive) · c85f2e9d72fb69ec · report
deeplabv3_resnet50 VainF/DeepLabV3Plus-Pytorch/network/modeling.py community (archive-listed) unverified MIT (permissive) · 0264c7923a17f00c · report

Tasks

2D Semantic SegmentationDichotomous Image SegmentationImage SegmentationSegmentationSemantic SegmentationThermal Image Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
2D Semantic Segmentation WildScenes DeepLabv3 (ResNet-50) mIoU 43.37 #4 of 5 Archive leaderboard report
2D Semantic Segmentation WildScenes DeepLabv3 (ResNet-50) mIoU (Env DA) 36.12 #4 of 5 Archive leaderboard report
2D Semantic Segmentation WildScenes DeepLabv3 (ResNet-50) mIoU (Temporal DA) 43.95 #4 of 5 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 DeeplabV3+ E-measure 0.772 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 DeeplabV3+ HCE 234 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 DeeplabV3+ MAE 0.102 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 DeeplabV3+ S-Measure 0.694 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 DeeplabV3+ max F-Measure 0.601 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE1 DeeplabV3+ weighted F-measure 0.506 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 DeeplabV3+ E-measure 0.813 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 DeeplabV3+ HCE 516 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 DeeplabV3+ MAE 0.105 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 DeeplabV3+ S-Measure 0.729 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 DeeplabV3+ max F-Measure 0.681 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE2 DeeplabV3+ weighted F-measure 0.587 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 DeeplabV3+ E-measure 0.833 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 DeeplabV3+ HCE 999 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 DeeplabV3+ MAE 0.102 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 DeeplabV3+ S-Measure 0.749 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 DeeplabV3+ max F-Measure 0.717 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE3 DeeplabV3+ weighted F-measure 0.623 #20 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 DeeplabV3+ E-measure 0.820 #19 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 DeeplabV3+ HCE 3709 #19 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 DeeplabV3+ MAE 0.111 #19 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 DeeplabV3+ S-Measure 0.744 #19 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 DeeplabV3+ max F-Measure 0.715 #19 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-TE4 DeeplabV3+ weighted F-measure 0.621 #19 of 22 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD DeeplabV3+ E-measure 0.796 #23 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD DeeplabV3+ HCE 1520 #23 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD DeeplabV3+ MAE 0.114 #23 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD DeeplabV3+ S-Measure 0.716 #23 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD DeeplabV3+ max F-Measure 0.660 #23 of 24 Archive leaderboard report
Dichotomous Image Segmentation DIS-VD DeeplabV3+ weighted F-measure 0.568 #23 of 24 Archive leaderboard report
Semantic Segmentation Cityscapes test DeepLabv3 (ResNet-101, coarse) Mean IoU (class) 81.3% #43 of 105 Archive leaderboard report
Semantic Segmentation Cityscapes val DeepLabv3 (Dilated-ResNet-101) mIoU 78.5% #61 of 99 Archive leaderboard report
Semantic Segmentation PASCAL VOC 2012 test DeepLabv3-JFT Mean IoU 86.9% #3 of 51 Archive leaderboard report
Semantic Segmentation PASCAL VOC 2012 val DeepLabv3-JFT mIoU 82.7% #6 of 29 Archive leaderboard report
Semantic Segmentation SELMA DeepLabV3 mIoU 70.7 #3 of 7 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

1x1 ConvolutionASPPAverage PoolingBatch NormalizationBottleneck Residual BlockConvolutionDeepLabv3Global Average PoolingKaiming InitializationMax PoolingPolynomial Rate DecayRandom Horizontal FlipRandom ScalingReLUResidual BlockResidual ConnectionSGD with MomentumSpatial Pyramid PoolingWeight Decay

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