pith. sign in

arxiv: 1907.10839 · v1 · pith:YCWDLMVUnew · submitted 2019-07-25 · 💻 cs.CV · cs.LG

Hard-Aware Fashion Attribute Classification

Yun Ye , Yixin Li , Bo Wu , Wei Zhang , Lingyu Duan , Tao Mei This is my paper

Pith reviewed 2026-05-24 16:34 UTC · model grok-4.3

classification 💻 cs.CV cs.LG
keywords fashion attribute classificationhard-aware backpropagationimbalanced dataGAN synthesissemi-supervised learningattribute recognitiondeep learning
0
0 comments X

The pith

A pipeline that adaptively emphasizes hard training samples and generates stable synthetic data for rare labels improves fashion attribute classification on imbalanced datasets.

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

The paper seeks to address extreme class imbalance in fashion attribute classification, where many attributes have very few positive examples. It introduces Hard-Aware BackPropagation to direct training effort toward difficult samples and a deactivated-output extension to semi-supervised GANs that produces complementary synthetic examples for the hardest attributes. These components together allow the model to learn more effectively from limited data without requiring extra labels. A reader would care because fashion attribute tasks underpin practical systems for search, recommendation, and trend analysis that routinely encounter the same imbalance problem.

Core claim

The authors claim that combining Hard-Aware BackPropagation, which adaptively weights training toward hard samples, with a Deact-modified semi-supervised GAN that deactivates outputs for synthetic samples to stabilize generation, yields higher accuracy on fashion attribute classification than prior methods when evaluated on large-scale imbalanced datasets.

What carries the argument

Hard-Aware BackPropagation (HABP) that re-weights gradients toward difficult examples, paired with the Deact modification to GAN training that prevents unstable outputs on synthetic complementary samples.

If this is right

  • HABP places higher training weight on hard samples during backpropagation.
  • Deact produces more stable synthetic samples specifically for attributes that lack sufficient real examples.
  • The combined pipeline outperforms prior state-of-the-art methods on large-scale fashion data.
  • All gains are obtained without any additional supervision or external labels.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same hard-sample emphasis and controlled synthesis steps could be tested on other imbalanced image classification problems outside fashion.
  • A direct distributional comparison between the generated complementary samples and real data could be added as an explicit validation step.
  • Removing either HABP or Deact individually on the same dataset would show how much each component contributes to the reported gains.

Load-bearing premise

The samples identified as hard by HABP are genuinely informative for improving generalization rather than noise, and the synthetic samples match the real data distribution closely enough to help training.

What would settle it

On the large-scale fashion dataset used in the paper, the full method produces lower or equal accuracy compared with existing state-of-the-art approaches that do not use HABP or Deact.

Figures

Figures reproduced from arXiv: 1907.10839 by Bo Wu, Lingyu Duan, Tao Mei, Wei Zhang, Yixin Li, Yun Ye.

Figure 1
Figure 1. Figure 1: Sample images and statistics of DeepFashion-C. Over 1/5 attributes have fewer than 100 positive samples. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Model predicted probabilites for positive samples vs. numbers of positive samples on DeepFashion-C. CE: [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Overall pipeline of the proposed method. HABP is calculated as the error probabilities weighted mean of [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Demonstration of generating complementary samples from a well trained GAN. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Architecture of MR-GAN. Decorrelation regularization is applied to the weights for projecting latent noise to [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Top-3 recall of HABP & FL under different learning rate. HABP demonstrates better stability and performance [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Training loss of attributes in the 1st epoch. HABP w/o DC w/ DC C o u n t 0 2×10 4 4×10 4 6×10 4 8×10 4 10 5 Cosine similarity −1 −0.5 0 0.5 1 [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Correlations between weights w/ and w/o decorrelation regularization [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Images generated using MR-GAN. (a) Upper: 224 × 224 samples generated using DeepFashion-C. Lower: 512 × 512 samples generated using CelebA-HQ. (b) Random samples w/ (right) and w/o (left) decorrelation regular￾ization loss. DeepFashion-C [6] CelebA-HQ [50] w/o DC 28.51 26.93 w/ DC 27.28 22.16 [PITH_FULL_IMAGE:figures/full_fig_p011_9.png] view at source ↗
read the original abstract

Fashion attribute classification is of great importance to many high-level tasks such as fashion item search, fashion trend analysis, fashion recommendation, etc. The task is challenging due to the extremely imbalanced data distribution, particularly the attributes with only a few positive samples. In this paper, we introduce a hard-aware pipeline to make full use of "hard" samples/attributes. We first propose Hard-Aware BackPropagation (HABP) to efficiently and adaptively focus on training "hard" data. Then for the identified hard labels, we propose to synthesize more complementary samples for training. To stabilize training, we extend semi-supervised GAN by directly deactivating outputs for synthetic complementary samples (Deact). In general, our method is more effective in addressing "hard" cases. HABP weights more on "hard" samples. For "hard" attributes with insufficient training data, Deact brings more stable synthetic samples for training and further improve the performance. Our method is verified on large scale fashion dataset, outperforming other state-of-the-art without any additional supervisions.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes Hard-Aware BackPropagation (HABP) to adaptively weight training toward hard samples/attributes in imbalanced fashion attribute classification, together with a deactivated semi-supervised GAN (Deact) that synthesizes complementary samples for those hard labels. The central claim is that the combined pipeline outperforms prior state-of-the-art methods on a large-scale fashion dataset without any additional supervision.

Significance. If the empirical gains are reproducible and the mechanisms are shown to be responsible rather than artifacts of data volume or regularization, the work would offer a practical route to improving tail-attribute performance in multi-label fashion tasks. The absence of extra supervision is a positive feature for deployment.

major comments (3)
  1. [Abstract] Abstract: the claim of outperformance supplies no metrics, baselines, ablation results, or error analysis, so the central empirical assertion cannot be evaluated from the provided text.
  2. [Method] Method (HABP description): the adaptive weighting is asserted to surface informative minority-attribute examples, yet no quantitative check (e.g., label-noise rate among selected hard samples or comparison against random oversampling) is supplied; if the selected samples are predominantly noisy, observed gains reduce to simple data-volume effects.
  3. [Method] Method (Deact-GAN): the claim that deactivating outputs for synthetic samples produces a marginal distribution sufficiently close to the real data to improve generalization on tail attributes lacks supporting evidence such as distribution-distance metrics, ablation of synthetic versus real oversampling, or visual inspection of generated samples.
minor comments (2)
  1. The acronyms HABP and Deact should be expanded on first use and the precise deactivation mechanism in the GAN loss should be written as an equation.
  2. Figure captions and axis labels in the experimental section require explicit mention of the dataset split and attribute frequency bins used for evaluation.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive feedback on clarifying empirical claims and strengthening evidence for the proposed methods. We address each major comment below with references to the manuscript content and indicate planned revisions where appropriate.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim of outperformance supplies no metrics, baselines, ablation results, or error analysis, so the central empirical assertion cannot be evaluated from the provided text.

    Authors: The abstract serves as a high-level summary of the contributions. Detailed quantitative results, including metrics, baselines, and ablation studies demonstrating outperformance on large-scale fashion datasets without extra supervision, are provided in Sections 4 and 5 of the manuscript. We will revise the abstract to incorporate key performance figures for better self-containment. revision: partial

  2. Referee: [Method] Method (HABP description): the adaptive weighting is asserted to surface informative minority-attribute examples, yet no quantitative check (e.g., label-noise rate among selected hard samples or comparison against random oversampling) is supplied; if the selected samples are predominantly noisy, observed gains reduce to simple data-volume effects.

    Authors: The manuscript demonstrates HABP's effectiveness via overall performance gains on hard attributes and qualitative examples of selected samples. We agree that explicit checks such as noise-rate analysis or comparisons to random oversampling would strengthen the argument against data-volume artifacts. We will add these quantitative analyses in the revision. revision: yes

  3. Referee: [Method] Method (Deact-GAN): the claim that deactivating outputs for synthetic samples produces a marginal distribution sufficiently close to the real data to improve generalization on tail attributes lacks supporting evidence such as distribution-distance metrics, ablation of synthetic versus real oversampling, or visual inspection of generated samples.

    Authors: The manuscript includes ablation studies comparing Deact to standard semi-supervised GAN, showing improved stability and performance on tail attributes. We acknowledge the absence of distribution-distance metrics, synthetic-vs-real oversampling ablations, and generated-sample visuals. We will incorporate these (e.g., FID scores and example images) in the revised version. revision: yes

Circularity Check

0 steps flagged

No circularity; empirical method with no derivation chain

full rationale

The paper proposes HABP for adaptive focus on hard samples and a Deact extension to semi-supervised GAN for synthesizing complementary data on imbalanced fashion attributes. The provided abstract and description contain no equations, no claimed first-principles derivations, and no predictions that reduce to fitted inputs or self-citations. The central claim rests on empirical outperformance on a large-scale dataset without extra supervision. No self-definitional, fitted-input, or self-citation patterns appear; the work is a self-contained empirical technique whose validity is independent of any internal reduction to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Review performed on abstract only; no explicit free parameters, axioms, or invented entities are stated beyond standard deep-learning assumptions of gradient-based optimization and GAN training stability.

pith-pipeline@v0.9.0 · 5715 in / 965 out tokens · 36967 ms · 2026-05-24T16:34:40.193341+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

62 extracted references · 62 canonical work pages · 5 internal anchors

  1. [1]

    Farhadi, I

    A. Farhadi, I. Endres, D. Hoiem, and D. Forsyth. Describing objects by their attributes. In 2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 1778–1785, June 2009

    work page 2009

  2. [2]

    K. Duan, D. Parikh, D. Crandall, and K. Grauman. Discovering localized attributes for fine-grained recognition. In 2012 IEEE Conference on Computer Vision and Pattern Recognition, pages 3474–3481, June 2012

    work page 2012

  3. [3]

    Siddiquie, R

    B. Siddiquie, R. S. Feris, and L. S. Davis. Image ranking and retrieval based on multi-attribute queries. In CVPR 2011, pages 801–808, June 2011

    work page 2011

  4. [4]

    B. Chen, Y . Chen, Y . Kuo, and W. H. Hsu. Scalable face image retrieval using attribute-enhanced sparse codewords. IEEE Transactions on Multimedia, 15(5):1163–1173, Aug 2013

    work page 2013

  5. [5]

    Hospedales, and Shaogang Gong

    Ryan Layne, Timothy M. Hospedales, and Shaogang Gong. Towards person identification and re-identification with attributes. In Andrea Fusiello, Vittorio Murino, and Rita Cucchiara, editors, Computer Vision – ECCV 2012. Workshops and Demonstrations, pages 402–412, Berlin, Heidelberg, 2012. Springer Berlin Heidelberg

    work page 2012

  6. [6]

    Z. Liu, P. Luo, S. Qiu, X. Wang, and X. Tang. Deepfashion: Powering robust clothes recognition and retrieval with rich annotations. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1096–1104, June 2016

    work page 2016

  7. [7]

    Simo-Serra, S

    E. Simo-Serra, S. Fidler, F. Moreno-Noguer, and R. Urtasun. Neuroaesthetics in fashion: Modeling the perception of fashionability. In 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 869–877, June 2015

    work page 2015

  8. [8]

    M. H. Kiapour, X. Han, S. Lazebnik, A. C. Berg, and T. L. Berg. Where to buy it: Matching street clothing photos in online shops. In 2015 IEEE International Conference on Computer Vision (ICCV), pages 3343–3351, Dec 2015

    work page 2015

  9. [9]

    Huang, R

    J. Huang, R. Feris, Q. Chen, and S. Yan. Cross-domain image retrieval with a dual attribute-aware ranking network. In 2015 IEEE International Conference on Computer Vision (ICCV), pages 1062–1070, Dec 2015

    work page 2015

  10. [10]

    Q. Chen, J. Huang, R. Feris, L. M. Brown, J. Dong, and S. Yan. Deep domain adaptation for describing people based on fine-grained clothing attributes. In 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 5315–5324, June 2015. 12

    work page 2015

  11. [11]

    S. Liu, Z. Song, G. Liu, C. Xu, H. Lu, and S. Yan. Street-to-shop: Cross-scenario clothing retrieval via parts alignment and auxiliary set. In 2012 IEEE Conference on Computer Vision and Pattern Recognition , pages 3330–3337, June 2012

    work page 2012

  12. [12]

    Towards better understanding the clothing fashion styles: A multimodal deep learning approach

    Yihui Ma, Jia Jia, Suping Zhou, Jingtian Fu, Yejun Liu, and Zijian Tong. Towards better understanding the clothing fashion styles: A multimodal deep learning approach. In Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence, AAAI’17, pages 38–44. AAAI Press, 2017

    work page 2017

  13. [13]

    StreetStyle: Exploring world-wide clothing styles from millions of photos

    Kevin Matzen, Kavita Bala, and Noah Snavely. Streetstyle: Exploring world-wide clothing styles from millions of photos. CoRR, abs/1706.01869, 2017

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  14. [14]

    Takagi, E

    M. Takagi, E. Simo-Serra, S. Iizuka, and H. Ishikawa. What makes a style: Experimental analysis of fashion prediction. In 2017 IEEE International Conference on Computer Vision Workshops (ICCVW), pages 2247–2253, Oct 2017

    work page 2017

  15. [15]

    Hsiao and K

    W. Hsiao and K. Grauman. Learning the latent “look”: Unsupervised discovery of a style-coherent embedding from fashion images. In 2017 IEEE International Conference on Computer Vision (ICCV), pages 4213–4222, Oct 2017

    work page 2017

  16. [16]

    Xintong Han, Zuxuan Wu, Yu-Gang Jiang, and Larry S. Davis. Learning fashion compatibility with bidirectional lstms. In Proceedings of the 25th ACM International Conference on Multimedia, MM ’17, pages 1078–1086, New York, NY , USA, 2017. ACM

    work page 2017

  17. [17]

    Hi, magic closet, tell me what to wear! In Proceedings of the 20th ACM International Conference on Multimedia, MM ’12, pages 619–628, New York, NY , USA, 2012

    Si Liu, Jiashi Feng, Zheng Song, Tianzhu Zhang, Hanqing Lu, Changsheng Xu, and Shuicheng Yan. Hi, magic closet, tell me what to wear! In Proceedings of the 20th ACM International Conference on Multimedia, MM ’12, pages 619–628, New York, NY , USA, 2012. ACM

    work page 2012

  18. [18]

    Neurostylist: Neural compatibility modeling for clothing matching

    Xuemeng Song, Fuli Feng, Jinhuan Liu, Zekun Li, Liqiang Nie, and Jun Ma. Neurostylist: Neural compatibility modeling for clothing matching. In Proceedings of the 25th ACM International Conference on Multimedia, MM ’17, pages 753–761, New York, NY , USA, 2017. ACM

    work page 2017

  19. [19]

    Creating capsule wardrobes from fashion images

    Wei-Lin Hsiao and Kristen Grauman. Creating capsule wardrobes from fashion images. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2018

    work page 2018

  20. [20]

    Hsu, and Jiebo Luo

    KuanTing Chen, Kezhen Chen, Peizhong Cong, Winston H. Hsu, and Jiebo Luo. Who are the devils wearing prada in new york city? In Proceedings of the 23rd ACM International Conference on Multimedia, MM ’15, pages 177–180, New York, NY , USA, 2015. ACM

    work page 2015

  21. [21]

    Fashion forward: Forecasting visual style in fashion

    Ziad Al-Halah, Rainer Stiefelhagen, and Kristen Grauman. Fashion forward: Forecasting visual style in fashion. In The IEEE International Conference on Computer Vision (ICCV), Oct 2017

    work page 2017

  22. [22]

    Learning from imbalanced data: open challenges and future directions

    Bartosz Krawczyk. Learning from imbalanced data: open challenges and future directions. Progress in Artificial Intelligence, 5(4):221–232, Nov 2016

    work page 2016

  23. [23]

    Chris Drummond, Robert C Holte, et al. C4. 5, class imbalance, and cost sensitivity: why under-sampling beats over-sampling. In Workshop on Learning from Imbalanced Datasets II, volume 11. Citeseer, 2003

    work page 2003

  24. [24]

    Chawla, Kevin W

    Nitesh V . Chawla, Kevin W. Bowyer, Lawrence O. Hall, and W. Philip Kegelmeyer. Smote: Synthetic minority over-sampling technique. J. Artif. Int. Res., 16(1):321–357, June 2002

    work page 2002

  25. [25]

    Kate McCarthy, Bibi Zabar, and Gary Weiss. Does cost-sensitive learning beat sampling for classifying rare classes? In Proceedings of the 1st international workshop on Utility-based data mining, UBDM ’05, pages 69–77, New York, NY , USA, 2005. ACM

    work page 2005

  26. [26]

    Rivera, María J

    Francisco Charte, Antonio J. Rivera, María J. del Jesus, and Francisco Herrera. Mlsmote. Know.-Based Syst., 89(C):385–397, November 2015

    work page 2015

  27. [27]

    Imbalanced multi-label learning for identifying antimicrobial peptides and their functional types

    Weizhong Lin and Dong Xu. Imbalanced multi-label learning for identifying antimicrobial peptides and their functional types. Bioinformatics, 32(24):3745–3752, 08 2016

    work page 2016

  28. [28]

    Hand, Carlos D

    Emily M. Hand, Carlos D. Castillo, and Rama Chellappa. Doing the best we can with what we have: Multi-label balancing with selective learning for attribute prediction. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, (AAAI-18), the 30th innovative Applications of Artificial Intelligence (IAAI-18), and the 8th AAAI Symposium on...

    work page 2018

  29. [29]

    Q. Dong, S. Gong, and X. Zhu. Imbalanced deep learning by minority class incremental rectification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 41(6):1367–1381, June 2019

    work page 2019

  30. [30]

    Training cost-sensitive neural networks with methods addressing the class imbalance problem

    Zhi-Hua Zhou and Xu-Ying Liu. Training cost-sensitive neural networks with methods addressing the class imbalance problem. IEEE Transactions on Knowledge and Data Engineering, 18(1):63–77, Jan 2006. 13

    work page 2006

  31. [31]

    S. H. Khan, M. Hayat, M. Bennamoun, F. A. Sohel, and R. Togneri. Cost-sensitive learning of deep feature representations from imbalanced data. IEEE Transactions on Neural Networks and Learning Systems, 29(8):3573– 3587, Aug 2018

    work page 2018

  32. [32]

    Cost-sensitive learning with neural networks

    Matjaz Kukar and Igor Kononenko. Cost-sensitive learning with neural networks. In Proceedings of the 13th European Conference on Artificial Intelligence (ECAI-98), pages 445–449. John Wiley & Sons, 1998

    work page 1998

  33. [33]

    Ling and Victor S

    Charles X. Ling and Victor S. Sheng. Cost-Sensitive Learning, pages 231–235. Springer US, Boston, MA, 2010

    work page 2010

  34. [34]

    Focal loss for dense object detection

    Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaiming He, and Piotr Dollar. Focal loss for dense object detection. In The IEEE International Conference on Computer Vision (ICCV), Oct 2017

    work page 2017

  35. [35]

    Training region-based object detectors with online hard example mining

    Abhinav Shrivastava, Abhinav Gupta, and Ross Girshick. Training region-based object detectors with online hard example mining. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2016

    work page 2016

  36. [36]

    Hard-aware deeply cascaded embedding

    Yuhui Yuan, Kuiyuan Yang, and Chao Zhang. Hard-aware deeply cascaded embedding. InThe IEEE International Conference on Computer Vision (ICCV), Oct 2017

    work page 2017

  37. [37]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In Z. Ghahramani, M. Welling, C. Cortes, N. D. Lawrence, and K. Q. Weinberger, editors, Advances in Neural Information Processing Systems 27, pages 2672–2680. Curran Associates, Inc., 2014

    work page 2014

  38. [38]

    Semi-Supervised Learning with Generative Adversarial Networks

    Augustus Odena. Semi-supervised learning with generative adversarial networks. CoRR, abs/1606.01583, 2016

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  39. [39]

    Unsupervised representation learning with deep convolutional generative adversarial networks

    Alec Radford, Luke Metz, and Soumith Chintala. Unsupervised representation learning with deep convolutional generative adversarial networks. In 4th International Conference on Learning Representations, ICLR 2016, San Juan, Puerto Rico, May 2-4, 2016, Conference Track Proceedings, 2016

    work page 2016

  40. [40]

    Towards Principled Methods for Training Generative Adversarial Networks

    Martín Arjovsky and Léon Bottou. Towards principled methods for training generative adversarial networks. CoRR, abs/1701.04862, 2017

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  41. [41]

    Describing clothing by semantic attributes

    Huizhong Chen, Andrew Gallagher, and Bernd Girod. Describing clothing by semantic attributes. In Andrew Fitzgibbon, Svetlana Lazebnik, Pietro Perona, Yoichi Sato, and Cordelia Schmid, editors, Computer Vision – ECCV 2012, pages 609–623, Berlin, Heidelberg, 2012. Springer Berlin Heidelberg

    work page 2012

  42. [42]

    Apparel classification with style

    Lukas Bossard, Matthias Dantone, Christian Leistner, Christian Wengert, Till Quack, and Luc Van Gool. Apparel classification with style. In Kyoung Mu Lee, Yasuyuki Matsushita, James M. Rehg, and Zhanyi Hu, editors, Computer Vision – ACCV 2012, pages 321–335, Berlin, Heidelberg, 2013. Springer Berlin Heidelberg

    work page 2012

  43. [43]

    Leveraging weakly annotated data for fashion image retrieval and label prediction

    Charles Corbiere, Hedi Ben-Younes, Alexandre Rame, and Charles Ollion. Leveraging weakly annotated data for fashion image retrieval and label prediction. In The IEEE International Conference on Computer Vision (ICCV) Workshops, Oct 2017

    work page 2017

  44. [44]

    Attentive fashion grammar network for fashion landmark detection and clothing category classification

    Wenguan Wang, Yuanlu Xu, Jianbing Shen, and Song-Chun Zhu. Attentive fashion grammar network for fashion landmark detection and clothing category classification. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2018

    work page 2018

  45. [45]

    Suykens and J

    J.A.K. Suykens and J. Vandewalle. Least squares support vector machine classifiers.Neural Processing Letters, 9(3):293–300, Jun 1999

    work page 1999

  46. [46]

    Facenet: A unified embedding for face recognition and clustering

    Florian Schroff, Dmitry Kalenichenko, and James Philbin. Facenet: A unified embedding for face recognition and clustering. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2015

    work page 2015

  47. [47]

    In Defense of the Triplet Loss for Person Re-Identification

    Alexander Hermans, Lucas Beyer, and Bastian Leibe. In defense of the triplet loss for person re-identification. CoRR, abs/1703.07737, 2017

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  48. [48]

    Kakadiaris

    Nikolaos Sarafianos, Xiang Xu, and Ioannis A. Kakadiaris. Deep imbalanced attribute classification using visual attention aggregation. In The European Conference on Computer Vision (ECCV), September 2018

    work page 2018

  49. [49]

    Deep generative image models using a laplacian pyramid of adversarial networks

    Emily L Denton, Soumith Chintala, arthur szlam, and Rob Fergus. Deep generative image models using a laplacian pyramid of adversarial networks. In C. Cortes, N. D. Lawrence, D. D. Lee, M. Sugiyama, and R. Garnett, editors, Advances in Neural Information Processing Systems 28, pages 1486–1494. Curran Associates, Inc., 2015

    work page 2015

  50. [50]

    Progressive growing of GANs for improved quality, stability, and variation

    Tero Karras, Timo Aila, Samuli Laine, and Jaakko Lehtinen. Progressive growing of GANs for improved quality, stability, and variation. In International Conference on Learning Representations, 2018

    work page 2018

  51. [51]

    MSG-GAN: multi-scale gradient GAN for stable image synthesis

    Animesh Karnewar, Oliver Wang, and Raghu Sesha Iyengar. MSG-GAN: multi-scale gradient GAN for stable image synthesis. CoRR, abs/1903.06048, 2019

    work page arXiv 1903

  52. [52]

    Improved techniques for training gans

    Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, Xi Chen, and Xi Chen. Improved techniques for training gans. In D. D. Lee, M. Sugiyama, U. V . Luxburg, I. Guyon, and R. Garnett, editors, Advances in Neural Information Processing Systems 29, pages 2234–2242. Curran Associates, Inc., 2016. 14

    work page 2016

  53. [53]

    Good semi-supervised learning that requires a bad gan

    Zihang Dai, Zhilin Yang, Fan Yang, William W Cohen, and Ruslan R Salakhutdinov. Good semi-supervised learning that requires a bad gan. In I. Guyon, U. V . Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors, Advances in Neural Information Processing Systems 30 , pages 6510–6520. Curran Associates, Inc., 2017

    work page 2017

  54. [54]

    Global versus localized generative adversarial nets

    Guo-Jun Qi, Liheng Zhang, Hao Hu, Marzieh Edraki, Jingdong Wang, and Xian-Sheng Hua. Global versus localized generative adversarial nets. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2018

    work page 2018

  55. [55]

    Large scale GAN training for high fidelity natural image synthesis

    Andrew Brock, Jeff Donahue, and Karen Simonyan. Large scale GAN training for high fidelity natural image synthesis. In International Conference on Learning Representations, 2019

    work page 2019

  56. [56]

    Conditional Generative Adversarial Nets

    Mehdi Mirza and Simon Osindero. Conditional generative adversarial nets. CoRR, abs/1411.1784, 2014

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  57. [57]

    Deep convolutional ranking for multilabel image annotation

    Yunchao Gong, Yangqing Jia, Thomas Leung, Alexander Toshev, and Sergey Ioffe. Deep convolutional ranking for multilabel image annotation. In 2nd International Conference on Learning Representations, ICLR 2014, Banff, AB, Canada, April 14-16, 2014, Conference Track Proceedings, 2014

    work page 2014

  58. [58]

    Very deep convolutional networks for large-scale image recognition

    Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recognition. In 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, 2015

    work page 2015

  59. [59]

    K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778, June 2016

    work page 2016

  60. [60]

    Lecun, L

    Y . Lecun, L. Bottou, Y . Bengio, and P. Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11):2278–2324, Nov 1998

    work page 1998

  61. [61]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, 2015

    work page 2015

  62. [62]

    Gans trained by a two time-scale update rule converge to a local nash equilibrium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. In I. Guyon, U. V . Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors, Advances in Neural Information Processing Systems 30, pages 6626–6637. Curran ...

    work page 2017