Pith. sign in

REVIEW 3 major objections 8 minor 35 references

Learning Semantic-Specific Graph Representation for Multi-Label Image Recognition

T0 review · 3 major / 8 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read SSGRL claims that grounding per-category features in word semantics and propagating them over a label co-occurrence graph achieves state-of-the-art multi-label recognition with only image-level labels.

desk verdict The semantic decoupling module is a real and cleanly ablated architectural contribution, but the paper's headline SOTA claim is not supported because the comparator set is stale and omits the contemporaneous graph-based methods. read the letter →

arxiv 1908.07325 v1 pith:OHP63QX4 submitted 2019-08-20 cs.CV

classification cs.CV
keywords multi-labelimagerecognitionsemanticdecouplingsemantic-specificrepresentationgraphneuralnetworklabelco-occurrenceattentionmechanismGloVewordembeddingsclassification
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

This paper sets out to establish that multi-label image recognition improves when each category is given its own visual feature, extracted under the guidance of the category's word meaning, and when those per-category features are then allowed to exchange information over a graph of label co-occurrence. The proposed Semantic-Specific Graph Representation Learning (SSGRL) framework trains end-to-end with only image-level labels, using GloVe embeddings to steer attention toward category-relevant regions and a gated graph propagation step to model mutual label dependencies. On the four benchmarks tested, it reports higher mean average precision than the compared methods, with the largest gain on COCO. If the claim is right, the practical payoff is a weakly supervised route to multi-label recognition that does not depend on part-level or bounding-box annotations.

What carries the argument

The load-bearing mechanism is the pairing of semantic guided attention with gated graph propagation. For each category $c$, the category name is embedded as $x_c$ and combined with each spatial location's image feature $f^I_{w,h}$ through low-rank bilinear pooling, $\tilde{f}^I_{c,wh} = P^\top(\tanh((U^\top f^I_{wh})\odot(V^\top x_c)))+b$, and a softmax over locations turns the fused scores into attention coefficients used to pool $f_c$. These per-category feature vectors initialize the nodes of a graph with adjacency $a_{cc'}$ equal to the training-set probability that $c'$ appears when $c$ appears; a GRU-like update aggregates neighbor messages from both directions, runs $T=3$ steps, and each node's final state is concatenated with its initial state before classification. The graph lets one category borrow contextual evidence from correlated categories while the attention mechanism keeps the visual evidence specific.

What would settle it

Measure the overlap between the semantic decoupling module's per-category attention maps and ground-truth object bounding boxes on COCO or PASCAL VOC; if high-confidence categories frequently attend to regions that contain no corresponding object, or if swapping the GloVe vectors for random fixed vectors leaves mAP essentially unchanged, then the reported improvements are not caused by semantic guidance.

Watch

Extended reading notes

Core claim

The paper's central claim is that a multi-label classifier can be cast as two cooperating modules: a semantic decoupling module that fuses each category's pretrained word embedding with the image feature map to compute a per-category attention map and pooled feature vector, and a semantic interaction module that arranges these vectors as nodes of a graph whose edge weights are empirical label co-occurrence frequencies and runs a gated propagation for three steps to contextualize each node. The contextualized state is concatenated with the original decoupled feature and fed to a per-category classifier. With this design, the authors report mAP of 95.0% on PASCAL VOC 2007, 94.8% on VOC 2012, 83.8% on COCO, and 36.6% on Visual Genome-500, corresponding to gains of 2.5, 2.6, 6.7, and 3.1 percentage points over the best compared methods; pre-training on COCO and fusing two input scales pushes VOC 2012 to 95.4%.

Load-bearing premise

The framework's gain rides on the assumption that word embeddings pretrained on text point to visually coherent regions for each category, so the attention module highlights the right places despite having no bounding-box or part-level supervision.

Editorial extensions

If this is right

  • A multi-label classifier can be trained end-to-end with only image-level labels and still learn category-specific regions, because semantic vectors stand in for part-level supervision.
  • Label interactions are captured simultaneously for all pairs through graph propagation, rather than sequentially as in RNN/LSTM decoders, so mutual dependencies are not forced through a linear order.
  • The framework scales to a large vocabulary without extra annotation: co-occurrence weights come from training labels and semantic vectors come from pretrained embeddings.
  • Pre-training on a larger dataset and fusing multi-scale predictions further raises accuracy, suggesting the learned representations transfer across multi-label benchmarks.

Reading between the lines

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

  • A quantitative test of the semantic decoupling claim would measure whether the attention maps align with ground-truth object boxes; the paper only shows qualitative examples, so attention-to-box overlap (IoU) is a natural next experiment.
  • Because the co-occurrence graph is built once from training statistics, it may encode dataset biases; a testable extension is to compare fixed co-occurrence edges with edges that are predicted per image or fine-tuned during training.
  • The GloVe vectors are frozen text embeddings; replacing them with visual-semantic vectors learned from the same image-label data would separate the contribution of pretrained textual knowledge from the contribution of the attention-plus-graph architecture.
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.

Referee Report

3 major / 8 minor

Summary. The paper proposes a Semantic-Specific Graph Representation Learning (SSGRL) framework for multi-label image recognition. The framework consists of two modules: a semantic decoupling module that uses GloVe word embeddings to guide a learned attention mechanism toward category-specific image regions, and a semantic interaction module that constructs a label co-occurrence graph from training-set statistics and propagates information through a gated graph neural network. The authors report experiments on Microsoft COCO, PASCAL VOC 2007/2012, and Visual Genome (VG-500), claiming sizable mAP improvements over prior state-of-the-art methods, with margins of 2.5%, 2.6%, 6.7%, and 3.1% respectively. An ablation study on COCO shows clear contributions from both modules, and the code is released.

Significance. If the stated claims are fully substantiated, this paper would make a solid contribution to multi-label image recognition: it demonstrates that text-derived category semantics can guide visual attention without part-level annotations, and that graph propagation over label co-occurrence can improve recognition performance. The ablation study in Table 5 is cleanly designed and provides convincing evidence that both the semantic decoupling and semantic interaction modules contribute to the overall gain. The release of code and models also supports reproducibility. However, the headline claim of outperforming 'current state-of-the-art methods' is not currently supported by the experimental comparison, which omits closely related graph-based multi-label methods and relies on uncontrolled comparisons from heterogeneous papers.

major comments (3)
  1. [Abstract; Section 4.2, Tables 1-4] The central claim of the paper, that SSGRL outperforms current state-of-the-art methods by 2.5%, 2.6%, 6.7%, and 3.1% on VOC 2007, VOC 2012, COCO, and VG-500, is not established because the comparison set is incomplete. The paper omits the graph-based multi-label recognition methods that are contemporaneous with this work, most notably ML-GCN (Chen, Zhao, et al., CVPR 2019), which uses a label co-occurrence graph with graph convolutional networks. Since the semantic interaction module in this paper is precisely a label-co-occurrence graph with GNN propagation, ML-GCN is a directly relevant baseline and should be included. Without such comparisons, the reported margins over 'current state-of-the-art' are unsubstantiated, and the actual contribution could be a modest increment over an already-published graph-based method.
  2. [Section 4.2, Tables 1-4] The comparisons against existing methods are not controlled. The numbers for competitors are taken from their original publications, which use different backbones, input scales, data augmentations, and pretraining protocols. For example, Table 1 compares with ResNet-SRN on COCO, but no evidence is given that the backbone, image size, or training schedule are aligned with the proposed setup; Table 4 states that ResNet-SRN is trained with its released code, yet the test split and training details are not aligned with the proposed method. To support the claimed margins, the authors should re-run key competitors under a unified training and evaluation protocol, or explicitly list the exact settings for each competitor and justify that the comparisons are fair.
  3. [Section 4.2, Tables 1-4] The paper reports no measure of variance or multi-seed results. Since stochastic optimization can induce non-negligible mAP variation, the exact margins quoted in the abstract are not supported without error bars or repeated runs. This is particularly important for the smaller claimed margins (e.g., 2.5% on VOC 2007), where run-to-run variation could be of similar magnitude. The authors should provide standard deviations over multiple training runs or at least a sensitivity analysis.
minor comments (8)
  1. [Section 3.3] In the graph construction paragraph, the node set is written as {v0, v2, ..., vC−1}, but it should be {v0, v1, ..., vC−1}.
  2. [Section 3.2 and Section 3.4] The symbol f_c is used both for the semantic-specific feature vector in Eq. (6) and for the classification function in Eq. (10), which is notationally confusing. Please use distinct symbols for these two quantities.
  3. [Section 4.2.3] The text refers to 'ResNet-SNR' but the method is 'ResNet-SRN'; please correct this typo.
  4. [Section 4.3.1] The explanation that 'Ours w/o SD performs slightly better than the baseline method, since it does not incur any additional information but increases the model complexity' is confusing; the clause about model complexity seems to contradict the observed improvement. Please rephrase.
  5. [Section 3.5] The statement 'we fix the parameters of the previous 92 convolutional layers in fcnn' is vague; please specify which layers or stages are frozen, and whether this choice was based on prior work or a design decision.
  6. [Section 4.2.3] The random split of 10,000 test images from Visual Genome should be released or the random seed specified to enable exact reproduction and fair comparison with future methods.
  7. [Section 4.2.1] The description of the top-3 evaluation protocol is ambiguous: the paper first says labels with top-3 highest scores are used, then says labels with probabilities lower than 0.5 are excluded. Please clarify how these two criteria interact.
  8. [Figure 4 and Section 4.3.1] The qualitative visualization in Figure 4 would be strengthened by a quantitative localization measure, such as using ground-truth bounding boxes to compute the precision/recall of the attended regions; this would address the concern that GloVe embeddings may not align with visual appearance for some categories.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the framework's components are not fitted to the test targets, and no prediction reduces to an input by construction.

full rationale

The paper's central claim is an empirical mAP comparison on held-out test labels of PASCAL VOC 2007/2012, COCO, and VG-500. The semantic decoupling module uses external GloVe word vectors (Eq. 2) and learns attention coefficients (Eqs. 3-5) from image features; the semantic interaction module builds a graph from training-set label co-occurrence (Section 3.3) and propagates features via a gated mechanism (Eqs. 7-9). None of these quantities is defined in terms of the test predictions or the reported mAP. The label co-occurrence matrix is a training-set prior, not the prediction target, and the final scores (Eq. 10) are computed from image-derived features, so there is no fitted parameter renamed as a prediction. The paper's citations to the authors' prior graph/attention works (e.g., [1,3,27]) are used as architectural inspiration and are not invoked as a uniqueness theorem or as evidence that the model must take this form. Any concern about omitted 2018/2019 graph-based baselines is a competitor-completeness issue, not circularity. The ablations (Table 5) further test the modules against a ResNet-101 baseline, giving independent evidence for the internal contribution claims.

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

No new physical or formal entities are introduced; the modules are implemented with standard neural-network components. The main external priors are GloVe embeddings and the training-set label co-occurrence matrix.

free parameters (4)
  • Label co-occurrence matrix A = Probabilities P(category c' | category c) estimated from training set labels
    Used to construct the graph in Section 3.3; this prior is estimated from the same training data used to fit the model and is not independently validated.
  • Graph propagation iterations T = 3
    Chosen by hand in Section 3.4; the paper does not report sensitivity to this hyperparameter.
  • Fusion dimensions d1 and d2 = 1024 and 1024
    Set in Section 3.4 for low-rank bilinear pooling; chosen without stated justification.
  • Hidden state dimension of the GNN = 2048
    Set in Section 3.4; hand-picked.
assumptions (4)
  • domain assumption GloVe word embeddings capture semantic relations between category names
    Invoked in Eq. (2); if false, semantic decoupling cannot guide attention correctly.
  • domain assumption ImageNet-pretrained ResNet-101 features transfer and lower layers can be fixed
    Section 3.5 fixes first 92 conv layers; standard but unverified for this task.
  • domain assumption Training-set label co-occurrence statistics generalize to the test set
    Section 3.3 computes A from training annotations; if test label distribution shifts, graph weights misguide propagation.
  • domain assumption Gated graph network message passing is an effective way to model interaction
    Borrowed from prior work [18,1,27,3]; no theoretical guarantee.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Semantic-Specific Graph Representation for Multi-Label Image Recognition." pith.science (2026). https://pith.science/paper/OHP63QX4

@misc{pith2026190807325,
  author       = {Pith},
  title        = {Pith review of: Learning Semantic-Specific Graph Representation for Multi-Label Image Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OHP63QX4}},
  note         = {Machine review of arXiv:1908.07325}
}
read the original abstract

Recognizing multiple labels of images is a practical and challenging task, and significant progress has been made by searching semantic-aware regions and modeling label dependency. However, current methods cannot locate the semantic regions accurately due to the lack of part-level supervision or semantic guidance. Moreover, they cannot fully explore the mutual interactions among the semantic regions and do not explicitly model the label co-occurrence. To address these issues, we propose a Semantic-Specific Graph Representation Learning (SSGRL) framework that consists of two crucial modules: 1) a semantic decoupling module that incorporates category semantics to guide learning semantic-specific representations and 2) a semantic interaction module that correlates these representations with a graph built on the statistical label co-occurrence and explores their interactions via a graph propagation mechanism. Extensive experiments on public benchmarks show that our SSGRL framework outperforms current state-of-the-art methods by a sizable margin, e.g. with an mAP improvement of 2.5%, 2.6%, 6.7%, and 3.1% on the PASCAL VOC 2007 & 2012, Microsoft-COCO and Visual Genome benchmarks, respectively. Our codes and models are available at https://github.com/HCPLab-SYSU/SSGRL.

Figures

Figures reproduced from arXiv: 1908.07325 by the authors.

Figure 1
Figure 1. Illustration of our Semantic-Specific Graph Representa [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of our Semantic-Specific Graph Representation Learning framework. Given an input image, we first feed it into [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The AP (in %) of each category of our proposed framework and the ResNet-101 baseline. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Several examples of input images (left), semantic feature [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 28 canonical work pages

  1. [1]

    Knowledge-embedded representation learning for fine-grained image recognition

    Tianshui Chen, Liang Lin, Riquan Chen, Yang Wu, and Xi- aonan Luo. Knowledge-embedded representation learning for fine-grained image recognition. In Proceedings of Inter- national Joint Conference on Artificial Intelligence (IJCAI), pages 627–634, 2018

  2. [2]

    Recurrent attentional reinforcement learning for multi-label image recognition

    Tianshui Chen, Zhouxia Wang, Guanbin Li, and Liang Lin. Recurrent attentional reinforcement learning for multi-label image recognition. In Proceedings of AAAI Conference on Artificial Intelligence, pages 6730–6737, 2018

  3. [3]

    Knowledge-embedded routing network for scene graph gen- eration

    Tianshui Chen, Weihao Yu, Riquan Chen, and Liang Lin. Knowledge-embedded routing network for scene graph gen- eration. In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019

  4. [4]

    A concept-based image retrieval system

    Tat-Seng Chua, Hung-Keng Pung, Guo-Jun Lu, and Hee-Sen Jong. A concept-based image retrieval system. In Proceed- ings of the International Conference on System Sciences, vol- ume 3, pages 590–598. IEEE, 1994

  5. [5]

    Imagenet: A large-scale hierarchical im- age database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical im- age database. In Proceedings of IEEE Conference on Com- puter Vision and Pattern Recognition (CVPR) , pages 248–

  6. [6]

    Subcategory-aware object clas- sification

    Jian Dong, Wei Xia, Qiang Chen, Jianshi Feng, Zhongyang Huang, and Shuicheng Yan. Subcategory-aware object clas- sification. In Proceedings of the IEEE Conference on Com- puter Vision and Pattern Recognition, pages 827–834, 2013

  7. [7]

    The pascal visual object classes (voc) challenge

    Mark Everingham, Luc Van Gool, Christopher KI Williams, John Winn, and Andrew Zisserman. The pascal visual object classes (voc) challenge. International Journal of Computer Vision, 88(2):303–338, 2010

  8. [8]

    Collective multi- label classification

    Nadia Ghamrawi and Andrew McCallum. Collective multi- label classification. In Proceedings of the 14th ACM Interna- tional Conference on Information and Knowledge Manage- ment, pages 195–200. ACM, 2005

Show all 35 references
  1. [9]

    Deep convolutional ranking for mul- tilabel image annotation

    Yunchao Gong, Yangqing Jia, Thomas Leung, Alexander To- shev, and Sergey Ioffe. Deep convolutional ranking for mul- tilabel image annotation. arXiv preprint arXiv:1312.4894 , 2013

  2. [10]

    Multi-label classification using conditional dependency networks

    Yuhong Guo and Suicheng Gu. Multi-label classification using conditional dependency networks. In Proceedings of International Joint Conference on Artificial Intelligence (IJ- CAI), page 1300, 2011

  3. [11]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778, 2016

  4. [12]

    Reinforced multi-label image classification by explor- ing curriculum

    Shiyi He, Chang Xu, Tianyu Guo, Chao Xu, and Dacheng Tao. Reinforced multi-label image classification by explor- ing curriculum. In Proceedings of AAAI Conference on Arti- ficial Intelligence, 2018

  5. [13]

    Long short-term memory

    Sepp Hochreiter and J ¨urgen Schmidhuber. Long short-term memory. Neural Computation, 9(8):1735–1780, 1997

  6. [14]

    Hadamard product for low-rank bilinear pooling

    Jin-Hwa Kim, Kyoung-Woon On, Woosang Lim, Jeonghee Kim, Jung-Woo Ha, and Byoung-Tak Zhang. Hadamard product for low-rank bilinear pooling. arXiv preprint arXiv:1610.04325, 2016

  7. [15]

    Adam: A method for stochastic optimization

    Diederik Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 , 2014

  8. [16]

    Visual genome: Connecting language and vision using crowdsourced dense image annotations

    Ranjay Krishna, Yuke Zhu, Oliver Groth, Justin Johnson, Kenji Hata, Joshua Kravitz, Stephanie Chen, Yannis Kalan- tidis, Li-Jia Li, David A Shamma, et al. Visual genome: Connecting language and vision using crowdsourced dense image annotations. International Journal of Compute...

  9. [17]

    Improving pairwise ranking for multi-label image classification

    Yuncheng Li, Yale Song, and Jiebo Luo. Improving pairwise ranking for multi-label image classification. In Proceedings of IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), pages 1837–1845. IEEE, 2017

  10. [18]

    Gated graph sequence neural networks

    Yujia Li, Daniel Tarlow, Marc Brockschmidt, and Richard Zemel. Gated graph sequence neural networks. arXiv preprint arXiv:1511.05493, 2015

  11. [19]

    Microsoft coco: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In Proceedings of European Conference on Computer Vision (ECCV), pages 740–755. Springer, 2014

  12. [20]

    Multi-label image classi- fication via knowledge distillation from weakly-supervised detection

    Yongcheng Liu, Lu Sheng, Jing Shao, Junjie Yan, Shim- ing Xiang, and Chunhong Pan. Multi-label image classi- fication via knowledge distillation from weakly-supervised detection. In Proceedings of the ACM International Confer- ence on Multimedia (ACM MM), 2018

  13. [21]

    Glove: Global vectors for word representation

    Jeffrey Pennington, Richard Socher, and Christopher Man- ning. Glove: Global vectors for word representation. In Pro- ceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1532–1543, 2014

  14. [22]

    Very deep convo- lutional networks for large-scale image recognition

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

  15. [23]

    Selective search for ob- ject recognition

    Jasper RR Uijlings, Koen EA Van De Sande, Theo Gev- ers, and Arnold WM Smeulders. Selective search for ob- ject recognition. International Journal of Computer Vision, 104(2):154–171, 2013

  16. [24]

    Cnn-rnn: A unified frame- work for multi-label image classification

    Jiang Wang, Yi Yang, Junhua Mao, Zhiheng Huang, Chang Huang, and Wei Xu. Cnn-rnn: A unified frame- work for multi-label image classification. arXiv preprint arXiv:1604.04573, 2016

  17. [25]

    Beyond object proposals: Random crop pool- ing for multi-label image recognition

    Meng Wang, Changzhi Luo, Richang Hong, Jinhui Tang, and Jiashi Feng. Beyond object proposals: Random crop pool- ing for multi-label image recognition. IEEE Transactions on Image Processing, 25(12):5678–5688, 2016

  18. [26]

    Multi-label image recognition by recurrently dis- covering attentional regions

    Zhouxia Wang, Tianshui Chen, Guanbin Li, Ruijia Xu, and Liang Lin. Multi-label image recognition by recurrently dis- covering attentional regions. In Proceedings of IEEE In- ternational Conference on Computer Vision (ICCV) , pages 464–472. IEEE, 2017

  19. [27]

    Deep reasoning with knowledge graph for social relationship understanding

    Zhouxia Wang, Tianshui Chen, Jimmy Ren, Weihao Yu, Hui Cheng, and Liang Lin. Deep reasoning with knowledge graph for social relationship understanding. In Proceedings of International Joint Conference on Artificial Intelligence (IJCAI), pages 2021–2018, 2018

  20. [28]

    Hcp: A flexible cnn framework for multi-label image classification

    Yunchao Wei, Wei Xia, Min Lin, Junshi Huang, Bingbing Ni, Jian Dong, Yao Zhao, and Shuicheng Yan. Hcp: A flexible cnn framework for multi-label image classification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(9):1901–1907, 2016

  21. [29]

    Correlative multi-label multi-instance im- age annotation

    Xiangyang Xue, Wei Zhang, Jie Zhang, Bin Wu, Jianping Fan, and Yao Lu. Correlative multi-label multi-instance im- age annotation. In Proceedings of International Conference on Computer Vision (ICCV), pages 651–658. IEEE, 2011

  22. [30]

    Exploit bounding box annota- tions for multi-label object recognition

    Hao Yang, Joey Tianyi Zhou, Yu Zhang, Bin-Bin Gao, Jianxin Wu, and Jianfei Cai. Exploit bounding box annota- tions for multi-label object recognition. InProceedings of the IEEE Conference on Computer Vision and Pattern Recogni- tion (CVPR), pages 280–288, 2016

  23. [31]

    Pinterest board recommendation for twitter users

    Xitong Yang, Yuncheng Li, and Jiebo Luo. Pinterest board recommendation for twitter users. In Proceedings of the ACM International Conference on Multimedia (ACM MM) , pages 963–966. ACM, 2015

  24. [32]

    Multi-label image classification with regional latent semantic dependencies

    Junjie Zhang, Qi Wu, Chunhua Shen, Jian Zhang, and Jian- feng Lu. Multi-label image classification with regional latent semantic dependencies. arXiv preprint arXiv:1612.01082 , 2016

  25. [33]

    Multi-label image classification with regional latent semantic dependencies

    Junjie Zhang, Qi Wu, Chunhua Shen, Jian Zhang, and Jian- feng Lu. Multi-label image classification with regional latent semantic dependencies. IEEE Transactions on Multimedia, 2018

  26. [34]

    Learning spatial regularization with image- level supervisions for multi-label image classification

    Feng Zhu, Hongsheng Li, Wanli Ouyang, Nenghai Yu, and Xiaogang Wang. Learning spatial regularization with image- level supervisions for multi-label image classification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 5513–5522, 2017

  27. [35]

    Lawrence Zitnick and Piotr Doll ´ar

    C. Lawrence Zitnick and Piotr Doll ´ar. Edge boxes: Locat- ing object proposals from edges. In Proceedings of Euro- pean Conference on Computer Vision (ECCV) , pages 391–

Pith tools

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