Semantic Correlation Promoted Shape-Variant Context for Segmentation

Ai Qun Liu; Bing Shuai; Gang Wang; Henghui Ding; Xudong Jiang

REVIEW 5 major objections 5 minor 77 references

Semantic Correlation Promoted Shape-Variant Context for Segmentation

T0 review · 5 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Semantic segmentation improves when each pixel draws context from a learned semantic-correlated shape rather than a fixed window.

desk verdict A new context-aggregation module with strong empirical results, but the key ablation does not isolate the proposed semantic-correlation mechanism. read the letter →

arxiv 1909.02651 v1 pith:DHKNGARI submitted 2019-09-05 cs.CV

Henghui Ding , Xudong Jiang , Bing Shuai , Ai Qun Liu , Gang Wang This is my paper · ORCID

classification cs.CV

keywords semanticsegmentationshape-variantcontextpairedconvolutioncorrelationmaskaggregationlabelingdenoisingsceneparsingdeepconvolutionalnetworks

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Semantic segmentation needs context: surrounding pixels often determine whether a pixel is a train, a lake, or a road. This paper claims that the usual approach—aggregating neighbors in a fixed square or a fixed global pool—is both wasteful and misleading, because the context that supports an object has a shape and scale tied to that object and its background. The authors propose to infer, for each pixel, a shape mask that marks which surrounding pixels are semantically correlated, then run a convolution whose receptive field is reshaped by that mask. They report that this shape-variant context, together with a labeling-denoising module that suppresses low-level noise, raises accuracy consistently on six segmentation benchmarks, and they show the shape mask is what drives the gain.

What carries the argument

The load-bearing object is the paired convolution plus Gaussian mapping that produces the shape mask. For each offset $(m,n)$ within a $K\times K$ window, the paired convolution applies one kernel at the target pixel and another at the neighbor; the difference $D_{m,n}^{i,j}$ is mapped by $\phi(a)=\exp(-a^2/\sigma^2)$ to a mask value that is high when the two features are judged correlated. The shape-variant convolution then reweights each tap of a standard location-invariant filter by $M_{m,n}^{i,j}$ (with depthwise-separable simplification for large kernels), so the effective receptive field at each pixel is customized by the input. The labeling-denoising module is secondary machinery: it computes category existence potentials from high-level score maps and subtracts learned penalties from lower-level score maps before fusion.

What would settle it

Measure, on a test set with dense instance or part annotations, the average shape-mask weight assigned to pixels of the same object or support category versus pixels of other categories inside the $K\times K$ window; the central claim would be falsified if same-category weights are not systematically higher, or if replacing the learned masks with random masks of the same statistics does not change accuracy.

Watch

Extended reading notes

Core claim

The central discovery is that context for classification can be made shape-variant in a differentiable, end-to-end way: a paired convolution estimates the semantic correlation between a target pixel and each neighbor, a Gaussian mapping turns the convolution discrepancy into a mask value $M_{m,n}^{i,j}=\phi(D_{m,n}^{i,j})$, and a shape-variant convolution multiplies its location-invariant filter weights by that mask before aggregating. The same mask values crop the fixed $K\times K$ kernel into different shapes and scales at different spatial positions and for different test images. On top of this, the labeling-denoising (LD) model uses high-level existence potentials to penalize lower-level score maps of categories unlikely to be in the image. The paper reports that this SVCNet outperforms previous state-of-the-art methods on COCO-Stuff, SIFT-Flow, CamVid, PASCAL-Person-Part, PASCAL-Context, and Cityscapes, and that in ablation SVC beats shape-fixed context at every kernel size tested.

Load-bearing premise

The central assumption is that pixels of the same object and its supporting context have correlated local appearances, so the paired convolution can learn to give them high mask weights; if that correlation is absent, the inferred masks will not match the intended context.

Editorial extensions

If this is right

Shape-variant context captures multi-shape and multi-scale information in a single convolution layer instead of requiring parallel or stacked fixed-size context branches.
Because the shape mask is inferred from the input at test time, the same trained filters adapt their receptive fields per image and per pixel, unlike standard convolutions or fixed pooling.
The labeling-denoising module reduces noisy low-level contributions by suppressing categories with low existence potential, and the paper's ablation shows larger gains on datasets with more classes (COCO-Stuff, 171 classes, versus PASCAL-Context, 59 classes).
If the reported numbers hold, the method improves mean IoU over prior state of the art on all six benchmarks, including large margins on CamVid, PASCAL-Person-Part, and PASCAL-Context.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The per-pixel shape mask is essentially a learned pairwise affinity field, so it could be reused as an attention or grouping signal for other dense prediction tasks such as depth estimation, boundary detection, or instance segmentation.
The SFC-versus-SVC comparison at matched kernel sizes suggests the improvement comes from shape variability rather than parameter count, which invites a direct test of whether the Gaussian mapping is essential or just a convenient monotone transform.
A natural extension is to apply the same mask-based reshaping to self-attention or large-kernel operators in vision backbones, where the masks could prune which positions contribute at each pixel and reduce the cost of global context.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Dear colleague,

You should know this paper proposes a paired-convolution module that learns, for each pixel, a per-location mask defining where to aggregate context in a convolution. The mask is generated by comparing a target pixel feature with neighboring features through a pair of convolutions, mapped by a Gaussian. That specific mechanism is new relative to ASPP, PSP, deformable conv, and PSANet; it is not just a rehash. The authors apply it within FCN-4s (ResNet-101) and report consistent SOTA gains on six public segmentation benchmarks, which is a real achievement if the numbers hold.

The paper’s strengths: the method is described clearly, and the ablation in Table 2 shows SVC outperforming SFC (constant mask) across all kernel sizes on PASCAL-Context, with the best SVC kernel larger than the best SFC kernel. That is consistent with the mask doing something beyond a fixed larger window. The labeling denoising module is separate and appears to give a modest but consistent gain, especially on many-class datasets. The visualizations in Fig. 7 are suggestive.

Where the paper is weaker: the headline claim attributes gains to semantic-correlation learning, but the Table 2 comparison does not isolate that mechanism. SVC includes the full paired-convolution branch (K^2 filter pairs) whereas SFC presumably drops it. So SVC has extra parameters and an input-dependent attention mechanism; the gap may reflect added capacity or generic attention, not masks aligned with object/context semantics. The sentence claiming “the performance gain is not simply brought by the increased number of parameters” only addresses fixed kernel size, not the added branch. The masks are never supervised or quantitatively evaluated against semantic boundaries; only selected visual examples are shown. There are also no error bars or repeated runs, and no code or models released, which matters for a paper whose main evidence is empirical. The term “shape-variant” is a bit generous: the mask is per-pixel and input-conditional, but it does not explicitly model object contours.

None of this makes the method obviously wrong. The central assumption—pixels of the same semantic object/context have correlated features—is plausible, and end-to-end training means the mask is free to learn something useful. The soft spots are addressable: a capacity-matched SFC (same extra parameters but no mask), a supervised attention or correlation analysis, and variance reporting.

Who is this for? Anyone working on context aggregation in dense prediction will find the paired-convolution idea worth engaging. It deserves a serious referee, but a referee should push on the attribution question before accepting the paper’s story.

Referee Report

5 major / 5 minor

Summary. This paper introduces SVCNet, a semantic segmentation network with three main components: a paired convolution that estimates per-pixel semantic correlation with neighboring positions, a shape-variant convolution that aggregates context using the inferred shape mask, and a labeling denoising model that suppresses low-level feature noise using high-level class-existence potentials. The method is evaluated on six public benchmarks (COCO-Stuff, SIFT-Flow, CamVid, PASCAL-Person-Part, PASCAL-Context, Cityscapes), reporting improvements over the listed baselines and an ablation study (Table 2) that compares shape-variant context (SVC) with shape-fixed context (SFC).

Significance. The idea of learning an input-dependent, per-pixel context mask from feature correlations is a potentially useful direction for semantic segmentation, and the architecture is clearly specified with all components end-to-end differentiable. The paper provides a structured ablation and covers multiple datasets. However, the central claim that performance gains come specifically from semantic-correlation-driven shape masks is not convincingly isolated from added model capacity, and the state-of-the-art comparisons are against a limited set of baselines. If the mechanism were rigorously verified, the contribution would be of interest to the segmentation community.

major comments (5)

[Table 2 and Section 4.1] The SVC-versus-SFC comparison does not isolate the effect of the learned shape mask: SFC (M=1) removes the entire paired-convolution side branch (Eqs. (1)-(2)), so it has fewer parameters and FLOPs than SVC. The claim in the table caption that the gain is 'not simply brought by the increased number of parameters' is therefore not supported by this comparison. Please provide a matched-parameter control (e.g., a version that retains the side branch but uses a fixed mask, or an SFC with proportionally more channels) to separate the effect of the mask from the effect of added capacity.
[Section 3.1 and Figure 7] The paper asserts that the inferred masks capture semantic correlation, but no quantitative evidence is provided; the visualizations in Fig. 7 are selected examples. Please evaluate the mask quality against ground-truth semantic relationships (e.g., compute the distribution of mask values for pixel pairs of same class vs different class, or a correlation with class-conditional co-occurrence) on a validation set. Without this, the evidence is consistent with the mask being a generic input-dependent attention pattern learned only from the segmentation loss.
[Section 4.3] The claim of 'new state-of-the-arts consistently on the six public segmentation datasets' is not substantiated because the comparison tables omit many strong baselines that were available at the time of submission (e.g., DeepLabv3+, PSANet, OCNet, DANet, HRNet). Please update the comparisons to contemporary methods under the same evaluation protocol, or restrict the claim accordingly.
[Section 4] No repeated runs or variance estimates are reported for any of the experiments, including the ablation study. Given the large reported gains (e.g., ~10 IoU points on PASCAL-Context in Table 1) and the small evaluation sets (SIFT-Flow, CamVid), the results may not be stable. Please report the mean and standard deviation over at least three runs, or otherwise justify the robustness of the reported numbers.
[Sections 3.1-3.2 and 4] The sensitivity to key hyperparameters (σ in Eq. (2), kernel size K in Table 2, and penalty threshold T in Eq. (6)) is not analyzed. The paper states that σ is not sensitive but provides no evidence; T is set to t, 2t, 4t without an ablation. Please include a sensitivity study for these parameters, especially T in the labeling denoising model, since they directly affect the mask values and the denoising behavior.

minor comments (5)

[Equation (1) and surrounding text] There is a typo in the text: it reads '(i−m, j −m)' when describing the features, but the equation uses '(i−m, j−n)'. Please correct the offset notation.
[Section 4.1] The text contains a typo: 'PASCAL-Conext' should be 'PASCAL-Context'.
[Section 4] The description of the depthwise-separable modification of Eq. (4) is underspecified. Please provide the exact equations or pseudocode so that the implementation of SVC with large kernels is reproducible.
[Figure 7] The gray-scale mask panels are difficult to interpret. Overlaying the masks on the input image with a color map and marking the center pixel would improve readability.
[Section 3.1 and Section 4] The paper does not report the runtime or memory overhead of the paired-convolution branch, which adds about 18·D·K^2 parameters. An efficiency comparison would help the reader assess the method's practical value.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SVCNet is trained end-to-end and evaluated on held-out benchmarks; the SFC/SVC ablation imbalance is an attribution confound, not a circular derivation.

full rationale

The paper's derivation chain is self-contained with respect to the six benchmark results: the shape mask M is defined from paired-convolution feature discrepancies (Eqs. 1-2) and then used to modulate convolution weights (Eqs. 3-4), with all parameters learned from the segmentation loss. No quantity that is called a prediction is a fitted parameter in disguise; the state-of-the-art claims are comparisons against external methods on standard held-out test sets. The only self-citations (e.g., CCL [18]) are related-work comparisons, not load-bearing justifications. The main concern, that Table 2's SFC-vs-SVC comparison adds both the mask mechanism and an entire learnable paired-convolution branch so the gain may reflect capacity rather than semantic correlation, is a legitimate experimental-control criticism but not circularity: SVC's result is not equivalent to its input by construction, and the mask mechanism could in principle fail. The explicit correlation assumption in Sec. 3.1 is stated as an assumption, not derived from the target result. Accordingly the circularity score is 0.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The method introduces architectural components rather than physical entities. Free parameters are limited to sigma, kernel size, and penalty threshold. The main assumptions concern feature correlation and feature robustness; these are domain assumptions that the experiments indirectly validate.

free parameters (3)

sigma in Gaussian mapping = 3
Set manually in Section 3.1; the paper claims insensitivity but provides no sensitivity analysis.
kernel size K for shape-variant convolution = 23
Selected as the best value on PASCAL-Context in the ablation of Table 2, then used for all datasets.
penalty threshold T in labeling denoising = t=1/C, 2t, 4t
Heuristic scaling with class count; no ablation on this threshold is reported.

assumptions (4)

domain assumption Feature appearances of pixels belonging to the same object and its context show strong correlation because they frequently coexist in training images.
This is the basis for the paired convolution to learn semantic correlation; it is stated as an observation/assumption in Section 3.1.
domain assumption Higher-level features are more robust to noise and encode what categories exist, while lower-level features provide spatial detail.
Used to justify the labeling denoising model in Section 3.2.
domain assumption A ResNet-101 pretrained on ImageNet provides suitable local features for segmentation.
The method depends on transfer learning from ImageNet; this is not evaluated in the paper.
standard math Standard convolution and Gaussian function are differentiable and correctly compute the described operations.
Assumed for the paired convolution and mask generation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Semantic Correlation Promoted Shape-Variant Context for Segmentation." pith.science (2026). https://pith.science/paper/DHKNGARI

@misc{pith2026190902651,
  author       = {Pith},
  title        = {Pith review of: Semantic Correlation Promoted Shape-Variant Context for Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DHKNGARI}},
  note         = {Machine review of arXiv:1909.02651}
}

read the original abstract

Context is essential for semantic segmentation. Due to the diverse shapes of objects and their complex layout in various scene images, the spatial scales and shapes of contexts for different objects have very large variation. It is thus ineffective or inefficient to aggregate various context information from a predefined fixed region. In this work, we propose to generate a scale- and shape-variant semantic mask for each pixel to confine its contextual region. To this end, we first propose a novel paired convolution to infer the semantic correlation of the pair and based on that to generate a shape mask. Using the inferred spatial scope of the contextual region, we propose a shape-variant convolution, of which the receptive field is controlled by the shape mask that varies with the appearance of input. In this way, the proposed network aggregates the context information of a pixel from its semantic-correlated region instead of a predefined fixed region. Furthermore, this work also proposes a labeling denoising model to reduce wrong predictions caused by the noisy low-level features. Without bells and whistles, the proposed segmentation network achieves new state-of-the-arts consistently on the six public segmentation datasets.

Figures

Figures reproduced from arXiv: 1909.02651 by the authors.

**Figure 3.** (Best viewed in color) A shape mask is inferred by the [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

**Figure 4.** Semantic correlation-dependent shape-variant context [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗

Figures from the paper (4 more)

**Figure 5.** Figure 5: In-context error, e.g., the first row, refers to incorrect [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

**Figure 6.** Figure 6: Network Architecture. We use ResNet-101 as our base [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

**Figure 7.** Figure 7: Four visual examples of the shape-variant masks [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

**Figure 8.** Figure 8: Qualitative segmentation examples on COCO-Stuff (1st [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

77 extracted references · 59 canonical work pages

[1]

Higher order conditional random ﬁelds in deep neural networks

Anurag Arnab, Sadeep Jayasumana, Shuai Zheng, and Philip HS Torr. Higher order conditional random ﬁelds in deep neural networks. In ECCV, 2016

work page 2016
[2]

Segnet: A deep convolutional encoder-decoder architecture for image segmentation

Vijay Badrinarayanan, Alex Kendall, and Roberto Cipolla. Segnet: A deep convolutional encoder-decoder architecture for image segmentation. TPAMI, 2017

work page 2017
[3]

Dense decoder shortcut connections for single-pass semantic segmentation

Piotr Bilinski and Victor Prisacariu. Dense decoder shortcut connections for single-pass semantic segmentation. In CVPR, 2018

work page 2018
[4]

Segmentation and recognition using structure from motion point clouds

Gabriel J Brostow, Jamie Shotton, Julien Fauqueur, and Roberto Cipolla. Segmentation and recognition using structure from motion point clouds. In ECCV, 2008

work page 2008
[5]

Loss max-pooling for semantic image segmentation

S Rota Bulo, Gerhard Neuhold, and Peter Kontschieder. Loss max-pooling for semantic image segmentation. In CVPR, 2017

work page 2017
[6]

Breuel, Federico Raue, and Marcus Liwicki

Wonmin Byeon, Thomas M. Breuel, Federico Raue, and Marcus Liwicki. Scene labeling with lstm recurrent neural networks. In CVPR, 2015

work page 2015
[7]

Coco- stuff: Thing and stuff classes in context

Holger Caesar, Jasper Uijlings, and Vittorio Ferrari. Coco- stuff: Thing and stuff classes in context. In CVPR, 2018

work page 2018
[8]

Semantic image segmentation with deep convolutional nets and fully connected crfs

Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille. Semantic image segmentation with deep convolutional nets and fully connected crfs. In ICLR, 2015

work page 2015

Show all 77 references

[9]

Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs

Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs. TPAMI, 2018

2018
[10]

Attention to scale: Scale-aware semantic image segmentation

Liang-Chieh Chen, Yi Yang, Jiang Wang, Wei Xu, and Alan L Yuille. Attention to scale: Scale-aware semantic image segmentation. In CVPR, 2016

2016
[11]

Encoder-decoder with atrous separable convolution for semantic image segmentation

Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. Encoder-decoder with atrous separable convolution for semantic image segmentation. arXiv:1802.02611, 2018

2018 arXiv
[12]

Detect what you can: Detecting and representing objects using holistic models and body parts

Xianjie Chen, Roozbeh Mottaghi, Xiaobai Liu, Sanja Fidler, Raquel Urtasun, and Alan Yuille. Detect what you can: Detecting and representing objects using holistic models and body parts. In CVPR, 2014

2014
[13]

Xception: Deep learning with depthwise separable convolutions

Francois Chollet. Xception: Deep learning with depthwise separable convolutions. In CVPR, 2017

2017
[14]

The cityscapes dataset for semantic urban scene understanding

Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benenson, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene understanding. In CVPR, 2016

2016
[15]

Boxsup: Exploiting bounding boxes to supervise convolutional networks for semantic segmentation

Jifeng Dai, Kaiming He, and Jian Sun. Boxsup: Exploiting bounding boxes to supervise convolutional networks for semantic segmentation. In ICCV, 2015

2015
[16]

Deformable convolutional networks

Jifeng Dai, Haozhi Qi, Yuwen Xiong, Yi Li, Guodong Zhang, Han Hu, and Yichen Wei. Deformable convolutional networks. In ICCV, 2017

2017
[17]

A classiﬁcation reﬁnement strategy for semantic segmentation

James W Davis, Christopher Menart, Muhammad Akbar, and Roman Ilin. A classiﬁcation reﬁnement strategy for semantic segmentation. arXiv:1801.07674, 2018

2018 arXiv
[18]

Context contrasted feature and gated multi- scale aggregation for scene segmentation

Henghui Ding, Xudong Jiang, Bing Shuai, Ai Qun Liu, and Gang Wang. Context contrasted feature and gated multi- scale aggregation for scene segmentation. In CVPR, 2018

2018
[19]

Learning hierarchical features for scene labeling

Clement Farabet, Camille Couprie, Laurent Najman, and Yann LeCun. Learning hierarchical features for scene labeling. TPAMI, 35(8), 2013

2013
[20]

A network structure to explicitly reduce confusion errors in semantic segmentation

Qichuan Geng, Xinyu Huang, Zhong Zhou, and Ruigang Yang. A network structure to explicitly reduce confusion errors in semantic segmentation. arXiv:1808.00313, 2018

2018 arXiv
[21]

Laplacian pyramid reconstruction and reﬁnement for semantic segmentation

Golnaz Ghiasi and Charless C Fowlkes. Laplacian pyramid reconstruction and reﬁnement for semantic segmentation. In ECCV, 2016

2016
[22]

Unpaired image captioning by language pivoting

Jiuxiang Gu, Shaﬁq Joty, Jianfei Cai, and Gang Wang. Unpaired image captioning by language pivoting. In ECCV, 2018

2018
[23]

Recent advances in convolutional neural networks

Jiuxiang Gu, Zhenhua Wang, Jason Kuen, Lianyang Ma, Amir Shahroudy, Bing Shuai, Ting Liu, Xingxing Wang, Gang Wang, Jianfei Cai, et al. Recent advances in convolutional neural networks. Pattern Recognition, 2018

2018
[24]

Scene graph generation with external knowledge and image reconstruction

Jiuxiang Gu, Handong Zhao, Zhe Lin, Sheng Li, Jianfei Cai, and Mingyang Ling. Scene graph generation with external knowledge and image reconstruction. In CVPR, 2019

2019
[25]

Hypercolumns for object segmentation and ﬁne-grained localization

Bharath Hariharan, Pablo Arbel ´aez, Ross Girshick, and Jitendra Malik. Hypercolumns for object segmentation and ﬁne-grained localization. In CVPR, 2015

2015
[26]

Deep residual learning for image recognition

Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In CVPR, 2016

2016
[27]

Labelbank: Revisiting global perspectives for semantic segmentation

Hexiang Hu, Zhiwei Deng, Guang-Tong Zhou, Fei Sha, and Greg Mori. Labelbank: Revisiting global perspectives for semantic segmentation. arXiv:1703.09891, 2017

2017 arXiv
[28]

Weinberger

Gao Huang, Zhuang Liu, Laurens van der Maaten, and Kilian Q. Weinberger. Densely connected convolutional networks. In CVPR, 2017

2017
[29]

Error correction for dense semantic image labeling

Yu-Hui Huang, Xu Jia, Stamatios Georgoulis, Tinne Tuytelaars, and Luc Van Gool. Error correction for dense semantic image labeling. arXiv:1712.03812, 2017

2017 arXiv
[30]

Scene parsing with global context embedding

Wei-Chih Hung, Yi-Hsuan Tsai, Xiaohui Shen, Zhe L Lin, Kalyan Sunkavalli, Xin Lu, and Ming-Hsuan Yang. Scene parsing with global context embedding. In ICCV, 2017

2017
[31]

Batch normalization: Accelerating deep network training by reducing internal covariate shift

Sergey Ioffe and Christian Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In ICML, 2015

2015
[32]

Gated feedback reﬁnement network for dense image labeling

Md Amirul Islam, Mrigank Rochan, Neil DB Bruce, and Yang Wang. Gated feedback reﬁnement network for dense image labeling. In CVPR, 2017

2017
[33]

The one hundred layers tiramisu: Fully convolutional densenets for semantic segmentation

Simon J ´egou, Michal Drozdzal, David Vazquez, Adriana Romero, and Yoshua Bengio. The one hundred layers tiramisu: Fully convolutional densenets for semantic segmentation. In CVPRW, 2017

2017
[34]

Adaptive afﬁnity ﬁelds for semantic segmentation

Tsung-Wei Ke, Jyh-Jing Hwang, Ziwei Liu, and Stella X Yu. Adaptive afﬁnity ﬁelds for semantic segmentation. InECCV, 2018

2018
[35]

Recurrent scene parsing with perspective understanding in the loop

Shu Kong and Charless C Fowlkes. Recurrent scene parsing with perspective understanding in the loop. In CVPR, 2018

2018
[36]

Efﬁcient inference in fully connected crfs with gaussian edge potentials

Philipp Kr ¨ahenb¨uhl and Vladlen Koltun. Efﬁcient inference in fully connected crfs with gaussian edge potentials. In NIPS, 2011

2011
[37]

Imagenet classiﬁcation with deep convolutional neural networks

Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classiﬁcation with deep convolutional neural networks. In NIPS, 2012

2012
[38]

Feature space optimization for semantic video segmentation

Abhijit Kundu, Vibhav Vineet, and Vladlen Koltun. Feature space optimization for semantic video segmentation. In CVPR, 2016. 9

2016
[39]

Semantic object parsing with graph lstm

Xiaodan Liang, Xiaohui Shen, Jiashi Feng, Liang Lin, and Shuicheng Yan. Semantic object parsing with graph lstm. In ECCV, 2016

2016
[40]

Semantic object parsing with local-global long short-term memory

Xiaodan Liang, Xiaohui Shen, Donglai Xiang, Jiashi Feng, Liang Lin, and Shuicheng Yan. Semantic object parsing with local-global long short-term memory. In CVPR, 2016

2016
[41]

Dynamic- structured semantic propagation network

Xiaodan Liang, Hongfei Zhou, and Eric Xing. Dynamic- structured semantic propagation network. In CVPR, 2018

2018
[42]

Multi-scale context intertwining for semantic segmentation

Di Lin, Yuanfeng Ji, Dani Lischinski, Daniel Cohen-Or, and Hui Huang. Multi-scale context intertwining for semantic segmentation. In ECCV, 2018

2018
[43]

Reﬁnenet: Multi-path reﬁnement networks for high- resolution semantic segmentation

Guosheng Lin, Anton Milan, Chunhua Shen, and Ian Reid. Reﬁnenet: Multi-path reﬁnement networks for high- resolution semantic segmentation. In CVPR, 2017

2017
[44]

G. Lin, C. Shen, A. van dan Hengel, and I. Reid. Efﬁcient piecewise training of deep structured models for semantic segmentation. In CVPR, 2016

2016
[45]

Nonparametric scene parsing: Label transfer via dense scene alignment

Ce Liu, Jenny Yuen, and Antonio Torralba. Nonparametric scene parsing: Label transfer via dense scene alignment. In CVPR, 2009

2009
[46]

Sift ﬂow: Dense correspondence across scenes and its applications

Ce Liu, Jenny Yuen, and Antonio Torralba. Sift ﬂow: Dense correspondence across scenes and its applications. TPAMI, 2011

2011
[47]

Feature boosting network for 3d pose estimation

Jun Liu, Henghui Ding, Amir Shahroudy, Ling-Yu Duan, Xudong Jiang, Gang Wang, and Alex Kot Chichung. Feature boosting network for 3d pose estimation. TPAMI, 2019

2019
[48]

Semantic image segmentation via deep parsing network

Ziwei Liu, Xiaoxiao Li, Ping Luo, Chen-Change Loy, and Xiaoou Tang. Semantic image segmentation via deep parsing network. In ICCV, 2015

2015
[49]

Learning markov clustering networks for scene text detection

Zichuan Liu, Guosheng Lin, Sheng Yang, Jiashi Feng, Weisi Lin, and Wang Ling Goh. Learning markov clustering networks for scene text detection. In CVPR, 2018

2018
[50]

Towards robust curve text detection with conditional spatial expansion

Zichuan Liu, Guosheng Lin, Sheng Yang, Fayao Liu, Weisi Lin, and Wang Ling Goh. Towards robust curve text detection with conditional spatial expansion. InCVPR, 2019

2019
[51]

Fully convolutional networks for semantic segmentation

Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmentation. In CVPR, 2015

2015
[52]

Are spatial and global constraints really necessary for segmentation? In ICCV, 2011

Aurelien Lucchi, Yunpeng Li, Xavier Boix, Kevin Smith, and Pascal Fua. Are spatial and global constraints really necessary for segmentation? In ICCV, 2011

2011
[53]

Feedforward semantic segmentation with zoom-out features

Mohammadreza Mostajabi, Payman Yadollahpour, and Gre- gory Shakhnarovich. Feedforward semantic segmentation with zoom-out features. In CVPR, 2015

2015
[54]

The role of context for object detection and semantic segmentation in the wild

Roozbeh Mottaghi, Xianjie Chen, Xiaobai Liu, Nam-Gyu Cho, Seong-Whan Lee, Sanja Fidler, Raquel Urtasun, and Alan Yuille. The role of context for object detection and semantic segmentation in the wild. In CVPR, 2014

2014
[55]

Learning deconvolution network for semantic segmentation

Hyeonwoo Noh, Seunghoon Hong, and Bohyung Han. Learning deconvolution network for semantic segmentation. In ICCV, 2015

2015
[56]

Large kernel matters – improve semantic segmentation by global convolutional network

Chao Peng, Xiangyu Zhang, Gang Yu, Guiming Luo, and Jian Sun. Large kernel matters – improve semantic segmentation by global convolutional network. In CVPR, 2017

2017
[57]

Recurrent convolu- tional neural networks for scene labeling

Pedro Pinheiro and Ronan Collobert. Recurrent convolu- tional neural networks for scene labeling. In ICML, 2014

2014
[58]

Imagenet large scale visual recognition challenge

Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. Imagenet large scale visual recognition challenge. IJCV, 2015

2015
[59]

Recursive context propagation network for semantic scene labeling

Abhishek Sharma, Oncel Tuzel, and Ming-Yu Liu. Recursive context propagation network for semantic scene labeling. In NIPS, 2014

2014
[60]

Fully convolutional networks for semantic segmentation

Evan Shelhamer, Jonathon Long, and Trevor Darrell. Fully convolutional networks for semantic segmentation. TPAMI, 2016

2016
[61]

Toward achieving robust low-level and high- level scene parsing

Bing Shuai, Henghui Ding, Ting Liu, Gang Wang, and Xudong Jiang. Toward achieving robust low-level and high- level scene parsing. TIP, 2019

2019
[62]

Scene segmentation with dag-recurrent neural networks

Bing Shuai, Zhen Zuo, Bing Wang, and Gang Wang. Scene segmentation with dag-recurrent neural networks. TPAMI, 2018

2018
[63]

Very deep convolutional networks for large-scale image recognition

Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556, 2014

2014 arXiv
[64]

Going deeper with convolutions

Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. Going deeper with convolutions. In CVPR, 2015

2015
[65]

Aanet: Attribute attention network for person re-identiﬁcation

Chiat-Pin Tay, Sharmili Roy, and Kim-Hui Yap. Aanet: Attribute attention network for person re-identiﬁcation. In CVPR, 2019

2019
[66]

Finding things: Image parsing with regions and per-exemplar detectors

Joseph Tighe and Svetlana Lazebnik. Finding things: Image parsing with regions and per-exemplar detectors. In CVPR, 2013

2013
[67]

Bridging category-level and instance-level semantic image segmentation

Zifeng Wu, Chunhua Shen, and Anton van den Hengel. Bridging category-level and instance-level semantic image segmentation. arXiv:1605.06885, 2016

2016 arXiv
[68]

Zoom better to see clearer: Human and object parsing with hierarchical auto-zoom net

Fangting Xia, Peng Wang, Liang-Chieh Chen, and Alan L Yuille. Zoom better to see clearer: Human and object parsing with hierarchical auto-zoom net. In ECCV, 2016

2016
[69]

Combining the best of convolutional layers and recurrent layers: A hybrid network for semantic segmentation

Zhicheng Yan, Hao Zhang, Yangqing Jia, Thomas Breuel, and Yizhou Yu. Combining the best of convolutional layers and recurrent layers: A hybrid network for semantic segmentation. arXiv:1603.04871, 2016

2016 arXiv
[70]

Context driven scene parsing with attention to rare classes

Jimei Yang, Brian Price, Scott Cohen, and Ming-Hsuan Yang. Context driven scene parsing with attention to rare classes. In CVPR, 2014

2014
[71]

Denseaspp for semantic segmentation in street scenes

Maoke Yang, Kun Yu, Chi Zhang, Zhiwei Li, and Kuiyuan Yang. Denseaspp for semantic segmentation in street scenes. In CVPR, 2018

2018
[72]

Learning a discriminative feature network for semantic segmentation

Changqian Yu, Jingbo Wang, Chao Peng, Changxin Gao, Gang Yu, and Nong Sang. Learning a discriminative feature network for semantic segmentation. In CVPR, 2018

2018
[73]

Multi-scale context aggregation by dilated convolutions

Fisher Yu and Vladlen Koltun. Multi-scale context aggregation by dilated convolutions. arXiv:1511.07122, 2015

2015 arXiv
[74]

Context encoding for semantic segmentation

Hang Zhang, Kristin Dana, Jianping Shi, Zhongyue Zhang, Xiaogang Wang, Ambrish Tyagi, and Amit Agrawal. Context encoding for semantic segmentation. In CVPR, 2018

2018
[75]

Pyramid scene parsing network

Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, and Jiaya Jia. Pyramid scene parsing network. In CVPR, 2017

2017
[76]

Psanet: Point-wise spatial attention network for scene parsing

Hengshuang Zhao, Yi Zhang, Shu Liu, Jianping Shi, Chen Change Loy, Dahua Lin, and Jiaya Jia. Psanet: Point-wise spatial attention network for scene parsing. In ECCV, 2018

2018
[77]

Conditional random ﬁelds as recurrent neural networks

Shuai Zheng, Sadeep Jayasumana, Bernardino Romera- Paredes, Vibhav Vineet, Zhizhong Su, Dalong Du, Chang Huang, and Philip HS Torr. Conditional random ﬁelds as recurrent neural networks. In ICCV, 2015. 10

2015

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Higher order conditional random ﬁelds in deep neural networks

Anurag Arnab, Sadeep Jayasumana, Shuai Zheng, and Philip HS Torr. Higher order conditional random ﬁelds in deep neural networks. In ECCV, 2016

work page 2016

[2] [2]

Segnet: A deep convolutional encoder-decoder architecture for image segmentation

Vijay Badrinarayanan, Alex Kendall, and Roberto Cipolla. Segnet: A deep convolutional encoder-decoder architecture for image segmentation. TPAMI, 2017

work page 2017

[3] [3]

Dense decoder shortcut connections for single-pass semantic segmentation

Piotr Bilinski and Victor Prisacariu. Dense decoder shortcut connections for single-pass semantic segmentation. In CVPR, 2018

work page 2018

[4] [4]

Segmentation and recognition using structure from motion point clouds

Gabriel J Brostow, Jamie Shotton, Julien Fauqueur, and Roberto Cipolla. Segmentation and recognition using structure from motion point clouds. In ECCV, 2008

work page 2008

[5] [5]

Loss max-pooling for semantic image segmentation

S Rota Bulo, Gerhard Neuhold, and Peter Kontschieder. Loss max-pooling for semantic image segmentation. In CVPR, 2017

work page 2017

[6] [6]

Breuel, Federico Raue, and Marcus Liwicki

Wonmin Byeon, Thomas M. Breuel, Federico Raue, and Marcus Liwicki. Scene labeling with lstm recurrent neural networks. In CVPR, 2015

work page 2015

[7] [7]

Coco- stuff: Thing and stuff classes in context

Holger Caesar, Jasper Uijlings, and Vittorio Ferrari. Coco- stuff: Thing and stuff classes in context. In CVPR, 2018

work page 2018

[8] [8]

Semantic image segmentation with deep convolutional nets and fully connected crfs

Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille. Semantic image segmentation with deep convolutional nets and fully connected crfs. In ICLR, 2015

work page 2015