Pith. sign in

REVIEW 3 major objections 6 minor 125 references

Robust Regression via Deep Negative Correlation Learning

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read A single convolutional network, trained with a negative-correlation penalty on K heads, can outperform both single-network and conventional-ensemble baselines on regression tasks while adding no parameters.

desk verdict Useful incremental extension of DNCL with real empirical gains on crowd counting and age estimation; Proposition 1 has a genuine proof gap and the evaluation needs code and seeds before the 'clear margin' claims hold. read the letter →

arxiv 1908.09066 v1 pith:6ULIGCCF submitted 2019-08-24 cs.CV

classification cs.CV
keywords deepregressionnegativecorrelationlearningensembleRademachercomplexitycrowdcountingageestimationimagesuper-resolution
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

The paper claims that a deep regression network can be made to behave like an ensemble of $K$ regressors at the parameter cost of one network, by adding a negative-correlation penalty to the usual squared-error loss. The per-member loss is $L_k = \frac12(G_k - Y)^2 - \lambda(G_k - \bar G)^2$, where $\bar G$ is the average of all $K$ outputs; the second term actively decorrelates the members so their errors cancel on unseen data. Across crowd counting, apparent personality analysis, age estimation, and image super-resolution, the authors report that this one change beats a single network, several robust losses, and a conventional ensemble of the same architecture. If correct, the result gives practitioners a cheap "ensemble for free" plug-in that works with any convolutional backbone.

What carries the argument

The load-bearing object is the amended loss $L_k = \frac12(G_k-Y)^2 - \lambda(G_k-\bar G)^2$ together with the bias-variance-covariance decomposition of ensemble error, $\mathbb{E}[(\bar G-Y)^2] = \frac{1}{K}\sum_k \mathbb{E}[(G_k-Y)^2] - \frac{1}{K}\sum_k \mathbb{E}[(G_k-\bar G)^2]$. The decomposition shows that penalizing each member's deviation from the ensemble average subtracts covariance from the total error, which is exactly the diversity mechanism. The implementation uses group convolution at the top layer to give each of the $K$ regressors a disjoint subset of the shared feature map, so the ensemble has no more parameters than a single network. The theoretical engine is Rademacher complexity: the paper proves Proposition 1, that the ensemble class has empirical Rademacher complexity $1/K$ times that of a conventional class, and uses it to argue that each subproblem fits random noise less easily.

What would settle it

Train the same backbone twice on one vision regression task, once with $\lambda=0$ (conventional ensemble) and once with $\lambda>0$, and also compute the empirical Rademacher complexity of the $K$-head group-convolution class on random labels; if the $\lambda>0$ ensemble does not beat the $\lambda=0$ ensemble on held-out error, or if the measured complexity ratio is not near $1/K$, the central claim fails.

Watch

Extended reading notes

Core claim

The central discovery is that negative correlation learning, previously used for shallow regression ensembles, transfers directly to deep convolutional regressors and yields a stronger ensemble without extra parameters. The authors define $K$ base regressors mounted on top of shared convolutional features, with the top feature map split into $K$ subsets by group convolution so each regressor sees a private slice. Each regressor minimizes its own squared error minus $\lambda$ times its squared deviation from the ensemble average; minimizing the deviation term drives the members to be negatively correlated on the training data, which through the bias-variance-covariance decomposition reduces the ensemble's mean squared error. The paper further claims that the empirical Rademacher complexity of the resulting ensemble class is $1/K$ that of a standard network, so each subproblem is easier to optimize, and supports the claim with experiments on four vision regression tasks.

Load-bearing premise

The theoretical guarantee rests on identifying the averaged ensemble of regressors with the single-network class inside the Rademacher complexity calculation; if that identification is not valid, the claim that each subproblem is $1/K$ as complex is unsupported, even if the trained method still works empirically.

Editorial extensions

If this is right

  • Swapping the loss in an existing regression CNN is a drop-in change: the method adds no parameters, so training and inference cost stay close to that of a single network.
  • The improvement should persist across regression tasks with continuous targets, since the loss is task-agnostic and the paper demonstrates it on density maps, trait scores, ages, and pixel-level residuals.
  • The optimal ensemble size $K$ is bounded by the feature dimension, because each member needs a private slice of the top feature map; the paper reports degradation at very large $K$ on crowd counting.
  • Combining NCL with other loss functions in the first term is possible, but the authors report it is usually weaker than using the squared-error first term, which guides how the loss should be deployed.
  • Because the method only changes the loss and the top layer, it is complementary to task-specific backbones and to other regularization strategies.

Reading between the lines

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

  • An extension the paper does not pursue: apply the same loss to other dense continuous regression tasks, such as monocular depth estimation or optical-flow prediction, where the top feature map naturally splits into $K$ groups.
  • A corollary of the paper's complexity claim is that the relative gain should grow as the training set shrinks, because Rademacher-complexity control matters most when data are scarce; a controlled study that varies training-set size could test this.
  • The paper does not measure the empirical Rademacher complexity of its actual trained models; computing it on random labels for the $K$-head group-convolution class would directly confirm or refute the claimed $1/K$ factor.
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 / 6 minor

Summary. The paper proposes Deep Negative Correlation Learning (DNCL), a deep regression ensemble in which K regressors share the lower convolutional feature extractor and are separated only by a group-convolution layer at the top. Each regressor is trained with the amended loss L_k = 1/2(G_k - Y)^2 - lambda (G_k - \tilde G)^2 (Eq. 11), where \tilde G is the ensemble average. The authors claim that, without extra network parameters, the method controls the bias-variance-covariance trade-off, that the averaged ensemble class has empirical Rademacher complexity 1/K of a conventional network (Proposition 1), and that DNCL outperforms L2, SmoothL1, and Tukey losses as well as several task-specific state-of-the-art methods on crowd counting, personality analysis, age estimation, and image super-resolution.

Significance. If verified, DNCL is an attractive and simple mechanism for obtaining ensemble-like gains without multiplying parameters: it is end-to-end trainable, complementary to backbone architecture choices, and tested on four rather diverse regression tasks. The paper is also honest in attributing the loss to the negative correlation learning literature and in disclosing the relation to the authors' CVPR 2018 paper. The theoretical claim, once repaired, would be a concise formal statement about the complexity of the averaged ensemble class. However, the empirical core is not yet fully established: the super-resolution gains are mostly within 0.1 dB, no error bars or multiple-seed statistics are reported, no code is released, and the 'conventional ensemble' baseline used in Table 8 is not an independently trained deep ensemble. The significance for TPAMI therefore depends on the authors addressing these verification gaps.

major comments (3)
  1. [Sec. 3.2.3, Eq. (22)] The proof of Proposition 1 is incomplete as written and should be rewritten. The equality inside the supremum where the sum of group-convolution outputs is replaced by a single conventional convolution must be justified explicitly: each W_k^Q is nonzero only on its channel group, so the sum over k is exactly a 1x1 convolution with a concatenated weight vector; then the factor 1/K follows by homogeneity of the empirical Rademacher complexity under scaling of the function class. The proof also uses Lemma 1, whose N=1 derivation drops the absolute value that appears in Definition 1, and Lemmas 2-3 omit a clear statement of the boundedness assumptions. In addition, the abstract claims that each sub-problem has reduced Rademacher complexity, whereas Proposition 1 as stated concerns the averaged ensemble class \tilde G; the relation between the two claims should be reconciled.
  2. [Sec. 4.5, Table 8] The baseline called 'Conventional Ensemble' in Table 8 is the same shared-backbone architecture with lambda = 0, not an ensemble of independently trained networks. This is clear from the text in Section 3.2.2, where the authors state that setting lambda = 0 achieves conventional ensemble learning. The comparison therefore isolates the effect of the correlation penalty but does not test the claim of outperforming conventional deep ensembles, which typically cost K times more compute. The authors should either include an independently trained K-network ensemble baseline, matching total parameter/compute budget where possible, or qualify the claim to say that DNCL outperforms a shared-backbone multi-head ensemble trained independently.
  3. [Sec. 4, Tables 1-7] The central empirical claim of universal superiority over L2, SmoothL1, Tukey, and task-specific baselines is not supported by significance information. No standard deviations, multiple-seed runs, or paired tests are reported; the only repeated-evaluation case, MORPH age estimation, averages five random partitions but still reports no variance. In Table 7 the DNCL gains over DRRN are at most 0.1 dB in every super-resolution row, and several entries are equal (e.g., Set5 x4: 31.7 vs 31.7; Set14 x2: 33.2 vs 33.2). Differences of this magnitude are within typical run-to-run variability in PSNR. The authors should report error bars over at least 3-5 seeds for the main comparisons, or perform paired significance tests, and ideally release code or per-run results to make the single-shot numbers reproducible.
minor comments (6)
  1. [Sec. 3.2.3, Definition 1] The word 'Radamacher' in Definition 1 is a typo and should read 'Rademacher'.
  2. [Sec. 3.2.1, Eq. (11)] In the first line of Eq. (11), the summation index j should run over j != k, not j != i, for consistency with the regressor index k.
  3. [Sec. 4.3] The text refers to the 'MPRPH' dataset; this should be MORPH, as in Table 6.
  4. [Sec. 2.2, Crowd Counting] The citation 'citewang2015deep' appears unresolved in the text and should be replaced with a proper reference.
  5. [Sec. 4.5, Effect of lambda and K] The authors state that lambda is set in [10^-3, 10^-2], but no ablation table or plot for lambda is reported; a sensitivity analysis for the central hyperparameter would strengthen the paper.
  6. [Table 7] The SSIM values in Table 7 appear to be scaled by 100, as noted in the text, but the table header simply says SSIM; the header should explicitly state the scaling factor.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation: the DNCL loss is constructed from an algebraic identity and evaluated on external benchmarks; the only self-citation is a disclosed preliminary version and is not load-bearing.

full rationale

The paper's central construction is not circular. Eqn. (10) is an algebraic identity for any K-predictor average, and the amended loss in Eqn. (11) is a per-sample objective designed to encourage negative correlation; it is not fitted to the benchmark results and no fitted constant is relabeled as a prediction. The bias-variance-covariance discussion and Eqn. (10) come from the external NCL literature [18], [19], not from the authors' own prior claims. Proposition 1's proof in Eqn. (22) contains a terse identification of the summed group-convolution class with the conventional convolution class; with unconstrained group kernel blocks and a shared backbone, the sum of K group-conv heads does span the full 1x1-conv class, so the 1/K Rademacher factor is defensible, though the boundedness assumption is omitted. This is a proof-completeness issue, not a self-referential reduction. The only self-citation is the disclosed CVPR 2018 preliminary version, which is used for provenance rather than as evidence for the theoretical or empirical claims. The 'Conventional Ensemble' baseline being a shared-backbone multi-head network with lambda=0 is a comparison-protocol concern, and the absence of code or error bars is a reproducibility concern; neither makes the derivation circular. Overall, no step in the paper reduces by construction to its own inputs.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central empirical claim rests on standard ensemble theory plus two task-specific design choices: the number of regressors K and the penalty lambda. The theoretical 1/K complexity result rests on an unproven identification of function classes.

free parameters (2)
  • lambda (correlation penalty) = not fixed; recommended range 1e-3 to 1e-2
    Controls the strength of the negative correlation penalty in Eqn. (11); chosen empirically per task, not derived.
  • K (ensemble size) = 64 for crowd counting, 8 for personality, 16 for super-resolution, 5 for age estimation
    Number of regressors; tuned per dataset; Table 11 shows severe performance collapse for K=128 on Shanghai Part A.
assumptions (5)
  • standard math Bias-variance-covariance decomposition of ensemble error (Eqns. 6-10)
    Standard decomposition used to motivate the loss; accepted.
  • standard math Lipschitz contraction property of Rademacher complexity (Lemma 1)
    Known result, but the proof in the paper contains a typo in the N=1 case.
  • domain assumption Boundedness assumption sup |G(x)-y| <= M (Lemmas 2 and 3)
    Needed for the Lipschitz constant pM^(p-1); the bound is assumed without discussion.
  • ad hoc to paper The ensemble class and the conventional class coincide under the supremum in Proposition 1 (Eqn. 22)
    This is the unproven identification that yields the 1/K factor; it is the weakest point of the theory.
  • domain assumption Group convolution partitions features into K subsets without losing information needed for the task
    The method's efficiency relies on each regressor seeing only a feature subset; the empirical K study shows this holds only for moderate K.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Robust Regression via Deep Negative Correlation Learning." pith.science (2026). https://pith.science/paper/6ULIGCCF

@misc{pith2026190809066,
  author       = {Pith},
  title        = {Pith review of: Robust Regression via Deep Negative Correlation Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6ULIGCCF}},
  note         = {Machine review of arXiv:1908.09066}
}
read the original abstract

Nonlinear regression has been extensively employed in many computer vision problems (e.g., crowd counting, age estimation, affective computing). Under the umbrella of deep learning, two common solutions exist i) transforming nonlinear regression to a robust loss function which is jointly optimizable with the deep convolutional network, and ii) utilizing ensemble of deep networks. Although some improved performance is achieved, the former may be lacking due to the intrinsic limitation of choosing a single hypothesis and the latter usually suffers from much larger computational complexity. To cope with those issues, we propose to regress via an efficient "divide and conquer" manner. The core of our approach is the generalization of negative correlation learning that has been shown, both theoretically and empirically, to work well for non-deep regression problems. Without extra parameters, the proposed method controls the bias-variance-covariance trade-off systematically and usually yields a deep regression ensemble where each base model is both "accurate" and "diversified". Moreover, we show that each sub-problem in the proposed method has less Rademacher Complexity and thus is easier to optimize. Extensive experiments on several diverse and challenging tasks including crowd counting, personality analysis, age estimation, and image super-resolution demonstrate the superiority over challenging baselines as well as the versatility of the proposed method.

Figures

Figures reproduced from arXiv: 1908.09066 by the authors.

Figure 1
Figure 1. Decision surfaces for classification of artificial spirals dataset for both (i) conventional ensemble learning and (ii) NCL learning. The shading [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Demonstration of the training process of conventional ensemble [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Details of the proposed DNCL. Regression is formulated as ensemble learning with the same amount of parameter as a single CNN. DNCL [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visual comparison for 4× super-resolution of different super-resolution results. Fig. 4b shows the ground-truth high resolution image cropped from the original image in Fig. 4a. TABLE 5 Comparison of the properties of the proposed method vs. the top teams in the 2016 C…
Figure 5
Figure 5. Figure 5: The Multi-Scale Blob module used in NCL. [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Visualization on the diversities with all 64 base models. The first [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

125 extracted references · 72 canonical work pages

  1. [1]

    Crowd counting with deep negative correlation learning,

    Z. Shi, L. Zhang, Y . Liu, X. Cao, Y . Ye, M.-M. Cheng, and G. Zheng, “Crowd counting with deep negative correlation learning,” in IEEE Conf. Comput. Vis. Pattern Recog., 2018, pp. 5382–5390

  2. [2]

    Deep age distribution learning for apparent age estimation,

    Z. Huo, X. Yang, C. Xing, Y . Zhou, P. Hou, J. Lv, and X. Geng, “Deep age distribution learning for apparent age estimation,” in IEEE Conf. Comput. Vis. Pattern Recog. Worksh., 2016, pp. 17–24

  3. [3]

    Chalearn LAP 2016: First round challenge on first impressions – dataset and results,

    V . Ponce-L ´opez, B. Chen, M. Oliu, C. Corneanu, A. Clap ´es, I. Guyon, X. Bar ´o, H. J. Escalante, and S. Escalera, “Chalearn LAP 2016: First round challenge on first impressions – dataset and results,” inEur. Conf. Comput. Vis., 2016

  4. [4]

    Image super-resolution via deep recursive residual network,

    Y . Tai, J. Yang, and X. Liu, “Image super-resolution via deep recursive residual network,” in IEEE Conf. Comput. Vis. Pattern Recog. , vol. 1, no. 4, 2017

  5. [5]

    Robust visual tracking using oblique random forests,

    L. Zhang, J. Varadarajan, P. N. Suganthan, N. Ahuja, and P. Moulin, “Robust visual tracking using oblique random forests,” in IEEE Conf. Comput. Vis. Pattern Recog., 2017

  6. [6]

    Fast R-CNN,

    R. Girshick, “Fast R-CNN,” in Int. Conf. Comput. Vis., 2015, pp. 1440– 1448

  7. [7]

    Robust estimation of a location parameter,

    P. J. Huber et al. , “Robust estimation of a location parameter,” The annals of mathematical statistics, vol. 35, no. 1, pp. 73–101, 1964

  8. [8]

    Robust optimization for deep regression,

    V . Belagiannis, C. Rupprecht, G. Carneiro, and N. Navab, “Robust optimization for deep regression,” in IEEE Conf. Comput. Vis. Pattern Recog., 2015, pp. 2830–2838

Show all 125 references
  1. [9]

    Robust statistics,

    P. J. Huber, “Robust statistics,” in International Encyclopedia of Statis- tical Science. Springer, 2011, pp. 1248–1251. IEEE TRANSACTIONS ON PATTERN ANAL YSIS AND MACHINE INTELLIGENCE 14

  2. [10]

    On the unification of line processes, outlier rejection, and robust statistics with applications in early vision,

    M. J. Black and A. Rangarajan, “On the unification of line processes, outlier rejection, and robust statistics with applications in early vision,” Int. J. Comput. Vis., vol. 19, no. 1, pp. 57–91, 1996

  3. [11]

    Ensemble methods in machine learning,

    T. G. Dietterich et al. , “Ensemble methods in machine learning,” Multiple Classifier Systems, vol. 1857, pp. 1–15, 2000

  4. [12]

    Random forests,

    L. Breiman, “Random forests,” Machine Learning, vol. 45, no. 1, pp. 5–32, 2001

  5. [13]

    Ensemble classification and regression-recent developments, applications and future directions,

    Y . Ren, L. Zhang, and P. N. Suganthan, “Ensemble classification and regression-recent developments, applications and future directions,” IEEE Comput. Intell. Mag., vol. 11, no. 1, pp. 41–53, 2016

  6. [14]

    Oblique random forest ensemble via least square estimation for time series forecasting,

    X. Qiu, L. Zhang, P. N. Suganthan, and G. A. Amaratunga, “Oblique random forest ensemble via least square estimation for time series forecasting,” Information Sciences, vol. 420, pp. 249–262, 2017

  7. [15]

    Learning to count with cnn boosting,

    E. Walach and L. Wolf, “Learning to count with cnn boosting,” in Eur. Conf. Comput. Vis. Springer, 2016, pp. 660–676

  8. [16]

    Incremental boosting convolutional neural network for facial action unit recognition,

    S. Han, Z. Meng, A.-S. Khan, and Y . Tong, “Incremental boosting convolutional neural network for facial action unit recognition,” inConf. Neural Inf. Process. Syst., 2016, pp. 109–117

  9. [17]

    Deep boosting,

    C. Cortes, M. Mohri, and U. Syed, “Deep boosting,” in Int. Conf. Mach. Learn., 2014, pp. 1179–1187

  10. [18]

    Evolutionary ensembles with negative correlation learning,

    Y . Liu, X. Yao, and T. Higuchi, “Evolutionary ensembles with negative correlation learning,” IEEE Trans. Evol. Comput. , vol. 4, no. 4, pp. 380–387, 2000

  11. [19]

    Managing diversity in regression ensembles,

    G. Brown, J. L. Wyatt, and P. Ti ˇno, “Managing diversity in regression ensembles,” JMLR, vol. 6, no. Sep, pp. 1621–1650, 2005

  12. [20]

    Faster r-cnn: Towards real-time object detection with region proposal networks,

    S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real-time object detection with region proposal networks,” in Conf. Neural Inf. Process. Syst., 2015, pp. 91–99

  13. [21]

    Deep convolutional network cascade for facial point detection,

    Y . Sun, X. Wang, and X. Tang, “Deep convolutional network cascade for facial point detection,” in IEEE Conf. Comput. Vis. Pattern Recog., 2013, pp. 3476–3483

  14. [22]

    Deeppose: Human pose estimation via deep neural networks,

    A. Toshev and C. Szegedy, “Deeppose: Human pose estimation via deep neural networks,” in IEEE Conf. Comput. Vis. Pattern Recog., 2014, pp. 1653–1660

  15. [23]

    Facial landmark detection by deep multi-task learning,

    Z. Zhang, P. Luo, C. C. Loy, and X. Tang, “Facial landmark detection by deep multi-task learning,” in Eur. Conf. Comput. Vis. Springer, 2014, pp. 94–108

  16. [24]

    Deep joint task learning for generic object extraction,

    X. Wang, L. Zhang, L. Lin, Z. Liang, and W. Zuo, “Deep joint task learning for generic object extraction,” in Conf. Neural Inf. Process. Syst., 2014, pp. 523–531

  17. [25]

    The random subspace method for constructing decision forests,

    I. Barandiaran, “The random subspace method for constructing decision forests,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 20, no. 8, 1998

  18. [26]

    Greedy function approximation: a gradient boosting machine,

    J. H. Friedman, “Greedy function approximation: a gradient boosting machine,” Annals of statistics, pp. 1189–1232, 2001

  19. [27]

    Deep regression forests for age estimation,

    W. Shen, Y . Guo, Y . Wang, K. Zhao, B. Wang, and A. L. Yuille, “Deep regression forests for age estimation,” in IEEE Conf. Comput. Vis. Pattern Recog., June 2018

  20. [28]

    Cross-scene crowd counting via deep convolutional neural networks,

    C. Zhang, H. Li, X. Wang, and X. Yang, “Cross-scene crowd counting via deep convolutional neural networks,” in IEEE Conf. Comput. Vis. Pattern Recog., 2015, pp. 833–841

  21. [29]

    Imagenet classification with deep convolutional neural networks,

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” in Conf. Neural Inf. Process. Syst., 2012, pp. 1097–1105

  22. [30]

    Crowdnet: a deep convolutional network for dense crowd counting,

    L. Boominathan, S. S. Kruthiventi, and R. V . Babu, “Crowdnet: a deep convolutional network for dense crowd counting,” in ACM Int. Conf. Multimedia. ACM, 2016, pp. 640–644

  23. [31]

    Multiscale multitask deep netvlad for crowd counting,

    Z. Shi, L. Zhang, Y . Sun, and Y . Ye, “Multiscale multitask deep netvlad for crowd counting,” IEEE Trans Ind. Informat. , vol. 14, no. 11, pp. 4953–4962, 2018

  24. [32]

    Single-image crowd counting via multi-column convolutional neural network,

    Y . Zhang, D. Zhou, S. Chen, S. Gao, and Y . Ma, “Single-image crowd counting via multi-column convolutional neural network,” in IEEE Conf. Comput. Vis. Pattern Recog., 2016, pp. 589–597

  25. [33]

    Towards perspective-free ob- ject counting with deep learning,

    D. Onoro-Rubio and R. J. L ´opez-Sastre, “Towards perspective-free ob- ject counting with deep learning,” in Eur. Conf. Comput. Vis. Springer, 2016, pp. 615–629

  26. [34]

    Switching convolutional neural network for crowd counting,

    D. B. Sam, S. Surya, and R. V . Babu, “Switching convolutional neural network for crowd counting,” in IEEE Conf. Comput. Vis. Pattern Recog., vol. 1, no. 3, 2017, p. 6

  27. [35]

    Counting in the wild,

    C. Arteta, V . Lempitsky, and A. Zisserman, “Counting in the wild,” in ECCV. Springer, 2016, pp. 483–498

  28. [36]

    Representation learning by learning to count,

    M. Noroozi, H. Pirsiavash, and P. Favaro, “Representation learning by learning to count,” in ICCV, 2017, pp. 5898–5906

  29. [37]

    Connecting meeting behavior with extraversion a systematic study,

    B. Lepri, R. Subramanian, K. Kalimeri, J. Staiano, F. Pianesi, and N. Sebe, “Connecting meeting behavior with extraversion a systematic study,”IEEE Trans Aff. Comput., vol. 3, no. 4, pp. 443–455, 2012

  30. [38]

    Facetube: pre- dicting personality from facial expressions of emotion in online con- versational video,

    J.-I. Biel, L. Teijeiro-Mosquera, and D. Gatica-Perez, “Facetube: pre- dicting personality from facial expressions of emotion in online con- versational video,” in ACM Int. Conf. Multimodal Interaction . ACM, 2012, pp. 53–56

  31. [39]

    Inferring mood in ubiquitous conversational video,

    D. Sanchez-Cortes, J.-I. Biel, S. Kumano, J. Yamato, K. Otsuka, and D. Gatica-Perez, “Inferring mood in ubiquitous conversational video,” in International Conference on Mobile and Ubiquitous Multimedia . ACM, 2013, p. 22

  32. [40]

    Inference of personality traits and affect schedule by analysis of spontaneous reactions to affective videos,

    M. K. Abadi, J. A. M. Correa, J. Wache, H. Yang, I. Patras, and N. Sebe, “Inference of personality traits and affect schedule by analysis of spontaneous reactions to affective videos,” in IEEE Conf. Autom. Face. Gest. Recog., vol. 1. IEEE, 2015, pp. 1–8

  33. [41]

    Please, tell me about yourself: automatic personality assessment using short self-presentations,

    L. M. Batrinca, N. Mana, B. Lepri, F. Pianesi, and N. Sebe, “Please, tell me about yourself: automatic personality assessment using short self-presentations,” in ACM Int. Conf. Multimodal Interaction . ACM, 2011, pp. 255–262

  34. [42]

    Deep bimodal regression for apparent personality analysis,

    C.-L. Zhang, H. Zhang, X.-S. Wei, and J. Wu, “Deep bimodal regression for apparent personality analysis,” inEur. Conf. Comput. Vis. Springer, 2016, pp. 311–324

  35. [43]

    Deep impression: Audiovisual deep residual networks for multimodal appar- ent personality trait recognition,

    Y . G ¨uc ¸l¨ut¨urk, U. G ¨uc ¸l¨u, M. A. van Gerven, and R. van Lier, “Deep impression: Audiovisual deep residual networks for multimodal appar- ent personality trait recognition,” in Eur. Conf. Comput. Vis. Springer, 2016, pp. 349–358

  36. [44]

    Bi-modal first impressions recognition using temporally ordered deep audio and stochastic visual features,

    A. Subramaniam, V . Patel, A. Mishra, P. Balasubramanian, and A. Mit- tal, “Bi-modal first impressions recognition using temporally ordered deep audio and stochastic visual features,” in Eur. Conf. Comput. Vis. Springer, 2016, pp. 337–348

  37. [45]

    Combining deep facial and ambient features for first impression estimation,

    F. G ¨urpınar, H. Kaya, and A. A. Salah, “Combining deep facial and ambient features for first impression estimation,” in Eur. Conf. Comput. Vis. Springer, 2016, pp. 372–385

  38. [46]

    First impressions: A survey on computer vision-based apparent personality trait analysis,

    J. Junior, C. Jacques, Y . G ¨uc ¸l¨ut¨urk, M. P ´erez, U. G ¨uc ¸l¨u, C. Andujar, X. Bar ´o, H. J. Escalante, I. Guyon, M. A. van Gerven et al. , “First impressions: A survey on computer vision-based apparent personality trait analysis,” arXiv preprint arXiv:1804.08046, 2018

  39. [47]

    Explain- ing first impressions: Modeling, recognizing, and explaining apparent personality from videos,

    H. J. Escalante, H. Kaya, A. A. Salah, S. Escalera, Y . Gucluturk, U. Guclu, X. Baro, I. Guyon, J. J. Junior, M. Madadi et al., “Explain- ing first impressions: Modeling, recognizing, and explaining apparent personality from videos,” arXiv preprint arXiv:1802.00745, 2018

  40. [48]

    Automatic age estimation based on facial aging patterns,

    X. Geng, Z.-H. Zhou, and K. Smith-Miles, “Automatic age estimation based on facial aging patterns,”IEEE Trans. Pattern Anal. Mach. Intell., vol. 29, no. 12, pp. 2234–2240, 2007

  41. [49]

    Human age estimation using bio-inspired features,

    G. Guo, G. Mu, Y . Fu, and T. S. Huang, “Human age estimation using bio-inspired features,” in IEEE Conf. Comput. Vis. Pattern Recog. IEEE, 2009, pp. 112–119

  42. [50]

    Simultaneous dimensionality reduction and human age estimation via kernel partial least squares regression,

    G. Guo and G. Mu, “Simultaneous dimensionality reduction and human age estimation via kernel partial least squares regression,” inIEEE Conf. Comput. Vis. Pattern Recog. IEEE, 2011, pp. 657–664

  43. [51]

    Demographic estimation from face images: Human vs. machine performance,

    H. Han, C. Otto, X. Liu, and A. K. Jain, “Demographic estimation from face images: Human vs. machine performance,” IEEE Trans. Pattern Anal. Mach. Intell., no. 6, pp. 1148–1161, 2015

  44. [52]

    Age regression from faces using random forests,

    A. Montillo and H. Ling, “Age regression from faces using random forests,” in IEEE Int. Conf. Image Process. IEEE, 2009, pp. 2465– 2468

  45. [53]

    Facial age estimation by learning from label distributions,

    X. Geng, C. Yin, and Z.-H. Zhou, “Facial age estimation by learning from label distributions,” IEEE Trans. Pattern Anal. Mach. Intell. , vol. 35, no. 10, pp. 2401–2412, 2013

  46. [54]

    Age estimation by multi-scale convolutional network,

    D. Yi, Z. Lei, and S. Z. Li, “Age estimation by multi-scale convolutional network,” in Asian Conf. Comput. Vis. Springer, 2014, pp. 144–158

  47. [55]

    Ordinal regression with multiple output cnn for age estimation,

    Z. Niu, M. Zhou, L. Wang, X. Gao, and G. Hua, “Ordinal regression with multiple output cnn for age estimation,” in IEEE Conf. Comput. Vis. Pattern Recog., 2016, pp. 4920–4928

  48. [56]

    Using ranking-cnn for age estimation,

    S. Chen, C. Zhang, M. Dong, J. Le, and M. Rao, “Using ranking-cnn for age estimation,” in IEEE Conf. Comput. Vis. Pattern Recog., 2017

  49. [57]

    Anchored regression networks applied to age estimation and super resolution,

    E. Agustsson, R. Timofte, and L. Van Gool, “Anchored regression networks applied to age estimation and super resolution,” in Int. Conf. Comput. Vis. IEEE, 2017, pp. 1652–1661

  50. [58]

    Deep cost- sensitive and order-preserving feature learning for cross-population age estimation,

    K. Li, J. Xing, C. Su, W. Hu, Y . Zhang, and S. Maybank, “Deep cost- sensitive and order-preserving feature learning for cross-population age estimation,” in Int. Conf. Comput. Vis., 2018

  51. [59]

    Deep expectation of real and apparent age from a single image without facial landmarks,

    R. Rothe, R. Timofte, and L. Van Gool, “Deep expectation of real and apparent age from a single image without facial landmarks,” Int. J. Comput. Vis., vol. 126, no. 2-4, pp. 144–157, 2018

  52. [60]

    Image super-resolution using deep convolutional networks,

    C. Dong, C. C. Loy, K. He, and X. Tang, “Image super-resolution using deep convolutional networks,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 38, no. 2, pp. 295–307, 2016

  53. [61]

    Deep people counting in extremely dense crowds,

    C. Wang, H. Zhang, L. Yang, S. Liu, and X. Cao, “Deep people counting in extremely dense crowds,” in ACM Int. Conf. Multimedia . ACM, 2015, pp. 1299–1302. IEEE TRANSACTIONS ON PATTERN ANAL YSIS AND MACHINE INTELLIGENCE 15

  54. [62]

    Real-time single image and video super- resolution using an efficient sub-pixel convolutional neural network,

    W. Shi, J. Caballero, F. Husz ´ar, J. Totz, A. P. Aitken, R. Bishop, D. Rueckert, and Z. Wang, “Real-time single image and video super- resolution using an efficient sub-pixel convolutional neural network,” in IEEE Conf. Comput. Vis. Pattern Recog., 2016, pp. 1874–1883

  55. [63]

    Collaborative representation cascade for single-image super- resolution,

    Y . Zhang, Y . Zhang, J. Zhang, D. Xu, Y . Fu, Y . Wang, X. Ji, and Q. Dai, “Collaborative representation cascade for single-image super- resolution,” IEEE Trans. Syst. Man Cy.-S., no. 99, pp. 1–16, 2017

  56. [64]

    Residual dense network for image super-resolution,

    Y . Zhang, Y . Tian, Y . Kong, B. Zhong, and Y . Fu, “Residual dense network for image super-resolution,” in IEEE Conf. Comput. Vis. Pattern Recog., 2018, pp. 2472–2481

  57. [65]

    Deeply-recursive convolutional network for image super-resolution,

    J. Kim, J. Kwon Lee, and K. Mu Lee, “Deeply-recursive convolutional network for image super-resolution,” in IEEE Conf. Comput. Vis. Pattern Recog., 2016, pp. 1637–1645

  58. [66]

    Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connec- tions,

    X. Mao, C. Shen, and Y .-B. Yang, “Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connec- tions,” in Conf. Neural Inf. Process. Syst., 2016, pp. 2802–2810

  59. [67]

    Image super-resolution via dual-state recurrent networks,

    W. Han, S. Chang, D. Liu, M. Yu, M. Witbrock, and T. S. Huang, “Image super-resolution via dual-state recurrent networks,” in IEEE Conf. Comput. Vis. Pattern Recog., 2018

  60. [68]

    Deep backprojection networks for super-resolution,

    M. Haris, G. Shakhnarovich, and N. Ukita, “Deep backprojection networks for super-resolution,” in IEEE Conf. Comput. Vis. Pattern Recog., 2018

  61. [69]

    Image super- resolution using very deep residual channel attention networks,

    Y . Zhang, K. Li, K. Li, L. Wang, B. Zhong, and Y . Fu, “Image super- resolution using very deep residual channel attention networks,” in ECCV, 2018, pp. 286–301

  62. [70]

    Residual non-local attention networks for image restoration,

    Y . Zhang, K. Li, K. Li, B. Zhong, and Y . Fu, “Residual non-local attention networks for image restoration,” in ICLR, 2019

  63. [71]

    Ntire 2017 challenge on single image super-resolution: Methods and results,

    R. Timofte, E. Agustsson, L. Van Gool, M.-H. Yang, and L. Zhang, “Ntire 2017 challenge on single image super-resolution: Methods and results,” inIEEE Conf. Comput. Vis. Pattern Recog., 2017, pp. 114–125

  64. [72]

    Do we need hundreds of classifiers to solve real world classification problems,

    M. Fern ´andez-Delgado, E. Cernadas, S. Barro, and D. Amorim, “Do we need hundreds of classifiers to solve real world classification problems,” JMLR, vol. 15, no. 1, pp. 3133–3181, 2014

  65. [73]

    Benchmarking ensemble classifiers with novel co-trained kernal ridge regression and random vector functional link ensembles [research frontier],

    L. Zhang and P. N. Suganthan, “Benchmarking ensemble classifiers with novel co-trained kernal ridge regression and random vector functional link ensembles [research frontier],” IEEE Comput. Intell. Mag., vol. 12, no. 4, pp. 61–72, 2017

  66. [74]

    Bagging predictors,

    L. Breiman, “Bagging predictors,” Machine learning, vol. 24, no. 2, pp. 123–140, 1996

  67. [75]

    Multi-scale context aggregation by dilated convolutions,

    F. Yu and V . Koltun, “Multi-scale context aggregation by dilated convolutions,”arXiv preprint arXiv:1511.07122, 2015

  68. [76]

    Sphereface: Deep hypersphere embedding for face recognition,

    W. Liu, Y . Wen, Z. Yu, M. Li, B. Raj, and L. Song, “Sphereface: Deep hypersphere embedding for face recognition,” in IEEE Conf. Comput. Vis. Pattern Recog., vol. 1, 2017, p. 1

  69. [77]

    Joint face detection and alignment using multitask cascaded convolutional networks,

    K. Zhang, Z. Zhang, Z. Li, and Y . Qiao, “Joint face detection and alignment using multitask cascaded convolutional networks,” IEEE Signal Processing Letters, vol. 23, no. 10, pp. 1499–1503, 2016

  70. [78]

    Temporal segment networks: Towards good practices for deep action recognition,

    L. Wang, Y . Xiong, Z. Wang, Y . Qiao, D. Lin, X. Tang, and L. Van Gool, “Temporal segment networks: Towards good practices for deep action recognition,” in Eur. Conf. Comput. Vis. Springer, 2016, pp. 20–36

  71. [79]

    Diversity creation methods: a survey and categorisation,

    G. Brown, J. Wyatt, R. Harris, and X. Yao, “Diversity creation methods: a survey and categorisation,”Information Fusion, vol. 6, no. 1, pp. 5–20, 2005

  72. [80]

    The impact of diversity on online ensemble learning in the presence of concept drift,

    L. L. Minku, A. P. White, and X. Yao, “The impact of diversity on online ensemble learning in the presence of concept drift,”IEEE TKDE, vol. 22, no. 5, pp. 730–742, 2009

  73. [81]

    Stochastic multiple choice learning for training diverse deep ensembles,

    S. Lee, S. P. S. Prakash, M. Cogswell, V . Ranjan, D. Crandall, and D. Batra, “Stochastic multiple choice learning for training diverse deep ensembles,” in NeurIPS, 2016, pp. 2119–2127

  74. [82]

    Fast decorrelated neural network ensembles with random weights,

    M. Alhamdoosh and D. Wang, “Fast decorrelated neural network ensembles with random weights,” Information Sciences , vol. 264, pp. 104–117, 2014

  75. [83]

    Ensembling neural networks: many could be better than all,

    Z.-H. Zhou, J. Wu, and W. Tang, “Ensembling neural networks: many could be better than all,” Artificial Intelligence, vol. 137, no. 1-2, pp. 239–263, 2002

  76. [84]

    Neural network ensembles,

    L. K. Hansen and P. Salamon, “Neural network ensembles,” IEEE TPAMI, no. 10, pp. 993–1001, 1990

  77. [85]

    Rademacher and gaussian complex- ities: Risk bounds and structural results,

    P. L. Bartlett and S. Mendelson, “Rademacher and gaussian complex- ities: Risk bounds and structural results,” JMLR, vol. 3, no. Nov, pp. 463–482, 2002

  78. [86]

    Empirical margin distributions and bounding the generalization error of combined classifiers,

    V . Koltchinskii, D. Panchenko et al. , “Empirical margin distributions and bounding the generalization error of combined classifiers,” The Annals of Statistics, vol. 30, no. 1, pp. 1–50, 2002

  79. [87]

    Caffe: Convolutional architecture for fast feature embedding,

    Y . Jia, E. Shelhamer, J. Donahue, S. Karayev, J. Long, R. Girshick, S. Guadarrama, and T. Darrell, “Caffe: Convolutional architecture for fast feature embedding,” CoRR, vol. abs/1408.5093, 2014

  80. [88]

    Multi-source multi-scale counting in extremely dense crowd images,

    H. Idrees, I. Saleemi, C. Seibert, and M. Shah, “Multi-source multi-scale counting in extremely dense crowd images,” inIEEE Conf. Comput. Vis. Pattern Recog., 2013, pp. 2547–2554

  81. [89]

    Stct: Sequentially training convolutional networks for visual tracking,

    L. Wang, W. Ouyang, X. Wang, and H. Lu, “Stct: Sequentially training convolutional networks for visual tracking,” inIEEE Conf. Comput. Vis. Pattern Recog., 2016, pp. 1373–1381

  82. [90]

    Density-aware person detection and tracking in crowds,

    M. Rodriguez, I. Laptev, J. Sivic, and J.-Y . Audibert, “Density-aware person detection and tracking in crowds,” in IEEE Conf. Comput. Vis. Pattern Recog. IEEE, 2011, pp. 2423–2430

  83. [91]

    Learning to count objects in images,

    V . Lempitsky and A. Zisserman, “Learning to count objects in images,” in Conf. Neural Inf. Process. Syst., 2010, pp. 1324–1332

  84. [92]

    Multi-scale convolutional neural networks for crowd counting,

    L. Zeng, X. Xu, B. Cai, S. Qiu, and T. Zhang, “Multi-scale convolutional neural networks for crowd counting,” arXiv preprint arXiv:1702.02359, 2017

  85. [93]

    Fully con- volutional crowd counting on highly congested scenes,

    M. Marsden, K. McGuiness, S. Little, and N. E. O’Connor, “Fully con- volutional crowd counting on highly congested scenes,” arXiv preprint arXiv:1612.00220, 2016

  86. [94]

    Decidenet: counting varying density crowds through attention guided detection and density estimation,

    J. Liu, C. Gao, D. Meng, and A. G. Hauptmann, “Decidenet: counting varying density crowds through attention guided detection and density estimation,” in IEEE Conf. Comput. Vis. Pattern Recog. , 2018, pp. 5197–5206

  87. [95]

    Morph: A longitudinal image database of normal adult age-progression,

    K. Ricanek and T. Tesafaye, “Morph: A longitudinal image database of normal adult age-progression,” in IEEE Conf. Autom. Face. Gest. Recog. IEEE, 2006, pp. 341–345

  88. [96]

    Overview of research on facial ageing using the fg-net ageing database,

    G. Panis, A. Lanitis, N. Tsapatsoulis, and T. F. Cootes, “Overview of research on facial ageing using the fg-net ageing database,” IET Biometrics, vol. 5, no. 2, pp. 37–46, 2016

  89. [97]

    Multi-task warped gaussian process for personalized age estimation,

    Y . Zhang and D.-Y . Yeung, “Multi-task warped gaussian process for personalized age estimation,” in IEEE Conf. Comput. Vis. Pattern Recog. IEEE, 2010, pp. 2622–2629

  90. [98]

    Cumulative attribute space for age and crowd density estimation,

    K. Chen, S. Gong, T. Xiang, and C. Change Loy, “Cumulative attribute space for age and crowd density estimation,” in IEEE Conf. Comput. Vis. Pattern Recog., 2013, pp. 2467–2474

  91. [99]

    Image-based human age estimation by manifold learning and locally adjusted robust regression,

    G. Guo, Y . Fu, C. R. Dyer, and T. S. Huang, “Image-based human age estimation by manifold learning and locally adjusted robust regression,” IEEE Trans. Image Process., vol. 17, no. 7, pp. 1178–1188, 2008

  92. [100]

    Ordinal hyperplanes ranker with cost sensitivities for age estimation,

    K.-Y . Chang, C.-S. Chen, and Y .-P. Hung, “Ordinal hyperplanes ranker with cost sensitivities for age estimation,” in IEEE Conf. Comput. Vis. Pattern Recog. IEEE, 2011, pp. 585–592

  93. [101]

    Deeply-learned feature for age estimation,

    X. Wang, R. Guo, and C. Kambhamettu, “Deeply-learned feature for age estimation,” in IEEE Winter Conference on Computer Vision . IEEE, 2015, pp. 534–541

  94. [102]

    A ranking approach for human ages estimation based on face images,

    K.-Y . Chang, C.-S. Chen, and Y .-P. Hung, “A ranking approach for human ages estimation based on face images,” in Int. Conf. Pattern Recog., 2010, pp. 3396–3399

  95. [103]

    Con- tourlet appearance model for facial age estimation,

    K. Luu, K. Seshadri, M. Savvides, T. D. Bui, and C. Y . Suen, “Con- tourlet appearance model for facial age estimation,” in International Joint Conference on Biometrics. IEEE, 2011

  96. [104]

    Some like it hot-visual guidance for preference prediction,

    R. Rothe, R. Timofte, and L. Van Gool, “Some like it hot-visual guidance for preference prediction,” inIEEE Conf. Comput. Vis. Pattern Recog., 2016, pp. 5553–5561

  97. [105]

    Label distribution learning forests,

    W. Shen, K. Zhao, Y . Guo, and A. L. Yuille, “Label distribution learning forests,” in Conf. Neural Inf. Process. Syst., 2017, pp. 834–843

  98. [106]

    Single image super-resolution from transformed self-exemplars,

    J.-B. Huang, A. Singh, and N. Ahuja, “Single image super-resolution from transformed self-exemplars,” in IEEE Conf. Comput. Vis. Pattern Recog., 2015, pp. 5197–5206

  99. [107]

    Fast and accurate image upscaling with super-resolution forests,

    S. Schulter, C. Leistner, and H. Bischof, “Fast and accurate image upscaling with super-resolution forests,” in IEEE Conf. Comput. Vis. Pattern Recog., 2015, pp. 3791–3799

  100. [108]

    Accurate image super-resolution using very deep convolutional networks,

    J. Kim, J. Kwon Lee, and K. Mu Lee, “Accurate image super-resolution using very deep convolutional networks,” in IEEE Conf. Comput. Vis. Pattern Recog., 2016, pp. 1646–1654

  101. [109]

    Rapid object detection using a boosted cascade of simple features,

    P. Viola and M. Jones, “Rapid object detection using a boosted cascade of simple features,” in IEEE Conf. Comput. Vis. Pattern Recog., vol. 1. IEEE, 2001, pp. I–I

  102. [110]

    Active appearance models,

    T. F. Cootes, G. J. Edwards, and C. J. Taylor, “Active appearance models,” IEEE Trans. Pattern Anal. Mach. Intell. , no. 6, pp. 681–685, 2001

  103. [111]

    Image super-resolution via sparse representation,

    J. Yang, J. Wright, T. S. Huang, and Y . Ma, “Image super-resolution via sparse representation,” IEEE Trans. Image Process., vol. 19, no. 11, pp. 2861–2873, 2010

  104. [112]

    A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics,

    D. Martin, C. Fowlkes, D. Tal, and J. Malik, “A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics,” in Int. Conf. Comput. Vis., vol. 2. IEEE, 2001, pp. 416–423. IEEE TRANSACTIONS ON PATTERN ...

  105. [113]

    Low-complexity single-image super-resolution based on nonnegative neighbor embedding,

    M. Bevilacqua, A. Roumy, C. Guillemot, and M. L. Alberi-Morel, “Low-complexity single-image super-resolution based on nonnegative neighbor embedding,” in Brit. Mach. Vis. Conf. BMV A press, 2012

  106. [114]

    On single image scale-up using sparse-representations,

    R. Zeyde, M. Elad, and M. Protter, “On single image scale-up using sparse-representations,” in International Conference on Curves and Surface. Springer, 2010, pp. 711–730

  107. [115]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE Trans. Image Process., vol. 13, no. 4, pp. 600–612, 2004

  108. [116]

    An information fidelity criterion for image quality assessment using natural scene statistics,

    H. R. Sheikh, A. C. Bovik, and G. De Veciana, “An information fidelity criterion for image quality assessment using natural scene statistics,” IEEE Trans. Image Process., vol. 14, no. 12, pp. 2117–2128, 2005

  109. [117]

    An analysis of diversity measures,

    E. K. Tang, P. N. Suganthan, and X. Yao, “An analysis of diversity measures,” Machine learning, vol. 65, no. 1, pp. 247–271, 2006

  110. [118]

    Rotation forest: A new classifier ensemble method,

    J. J. Rodriguez, L. I. Kuncheva, and C. J. Alonso, “Rotation forest: A new classifier ensemble method,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 28, no. 10, pp. 1619–1630, 2006

  111. [119]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in IEEE Conf. Comput. Vis. Pattern Recog. , 2016, pp. 770–778

  112. [120]

    Wide Residual Networks,

    S. Zagoruyko and N. Komodakis, “Wide Residual Networks,” in Brit. Mach. Vis. Conf., 2016

  113. [121]

    Rank-based pooling for deep convolutional neural networks,

    Z. Shi, Y . Ye, and Y . Wu, “Rank-based pooling for deep convolutional neural networks,” Neural Networks, vol. 83, pp. 21–31, 2016

  114. [122]

    Richer convolutional features for edge detection,

    Y . Liu, M. Cheng, X. Hu, J. Bian, L. Zhang, X. Bai, and J. Tang, “Richer convolutional features for edge detection,”IEEE Trans. Pattern Anal. Mach. Intell. doi=10.1109/TPAMI.2018.2878849, pp. 1–1, 2018

  115. [123]

    Bag of tricks for convolutional neural network,

    T. He, Z. Zhang, H. Zhang, Z. Zhang, J. Xie, and M. Li, “Bag of tricks for convolutional neural network,” arXiv preprint arXiv:1812.01187 , 2018

  116. [124]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” CoRR, vol. abs/1409.1556, 2015. Le Zhang received the B.Eng degree from Uni- versity of Electronic Science and Technology Of China in 2011. He received his M.Sc and Ph.D.degree ...

  117. [2016]

    Jia-Wang Bian is a PhD student at the Uni- versity of Adelaide and an Associated PhD re- searcher with the Australian Centre for Robotic Vision (ACRV)

    His research interests include computer vision and machine learning. Jia-Wang Bian is a PhD student at the Uni- versity of Adelaide and an Associated PhD re- searcher with the Australian Centre for Robotic Vision (ACRV). He is advised by Prof. Ian Reid and Prof. Chunhua Shen. ...

Pith tools

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