Pith. sign in

REVIEW 2 major objections 5 minor 47 references

Needles in Haystacks: On Classifying Tiny Objects in Large Images

T0 review · 2 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper claims that image-level CNN classifiers have a minimum object size for a given training budget, and that the training data needed rises steeply as the object-to-image ratio shrinks.

desk verdict Useful testbed and a lot of honest empiricism, but the headline O2I scaling law is confounded by clutter count in nMNIST. read the letter →

arxiv 1908.06037 v2 pith:RQJ6Y3OP submitted 2019-08-16 cs.CV cs.LG

Nick Pawlowski , Suvrat Bhooshan , Nicolas Ballas , Francesco Ciompi , Ben Glocker , Michal Drozdzal This is my paper · ORCID
classification cs.CVcs.LG
keywords tinyobjectclassificationobject-to-imageratiosignal-to-noiseCNNgeneralizationdatasetscalingglobalpoolinghistopathologyreceptivefield
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 tries to establish that convolutional neural networks trained only with image-level labels have a hard lower limit on how small the class-defining object can be relative to the whole image, and that this limit moves when the training set changes. On a controlled testbed that varies the object-to-image ratio from roughly 19% down to 0.075%, networks generalize well at larger ratios and then fall to chance at a threshold; the smaller the object, the more training data is needed, and the required amount grows rapidly as the ratio shrinks. The paper argues this behavior appears in both a simple digit-classification task and a histopathology task, so it is not an artifact of one domain. If the claim holds, it matters for medical and remote-sensing applications where the signal can occupy a tiny fraction of the image and pixel-level labels are costly.

What carries the argument

The central object is the object-to-image (O2I) ratio, defined as the area of the class-relevant object divided by the area of the image. The argument is carried by two controlled datasets: a synthetic cluttered-digit set the paper calls nMNIST, which places a fixed 28-by-28 digit on canvases from 64-by-64 to 1024-by-1024 pixels while scaling the number of distractor digits from 2 to 400 to keep clutter density roughly constant, and a histopathology crop set the paper calls nCAMELYON, derived from pixel-level lesion annotations. Around this testbed the paper varies three network properties independently: receptive field size, model capacity, and the global pooling operation, using a convolutional backbone that permits explicit receptive-field control.

What would settle it

Re-run the 1024-by-1024 digit experiment with the same 0.075% object-to-image ratio but with, say, 10 clutter digits instead of 400; if the network then classifies above chance, the O2I limit is really a clutter-count limit, and the paper's headline scaling would not transfer to constant-clutter settings.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that there is a minimum object-to-image ratio for a given CNN training budget, and that below this ratio image-level classifiers do not beat random guessing on held-out data. More data lowers the floor, but only at a steep cost: the number of training examples needed to reach a fixed validation accuracy rises quickly as the object-to-image ratio drops, so the gains from simply collecting more data diminish sharply. The paper also reports that at small object-to-image ratios higher-capacity models generalize better, that a receptive field matched to or larger than the object helps when the object size is known, and that global max pooling is the most reliable choice for very small signals, while pooling choice matters little at larger ratios.

Load-bearing premise

The load-bearing premise is that the failure threshold is controlled by the object's share of the image, but the main testbed changes the number of distractor digits together with the canvas size, so the claimed scaling could be driven by crowding rather than by object size.

Editorial extensions

If this is right

  • At any fixed training-set size there is a smallest object size for which an image-level CNN can classify better than chance, so practitioners working with very small targets should expect random performance unless they can grow the dataset.
  • Dataset collection effort must grow much faster than the reduction in object size: a small further decrease in the object-to-image ratio demands a large increase in training examples to hold accuracy.
  • Raising model capacity improves low-signal generalization, but the benefit shrinks when the training set is small, so capacity alone is not a substitute for data.
  • If an approximate object size is known, setting the receptive field to cover it is a useful inductive bias; the wrong pooling choice can hurt optimization mainly in the very low signal-to-noise regime.

Reading between the lines

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

  • Because the digit testbed raises canvas size and clutter count together, the inverse-O2I scaling may actually measure how well the network can ignore many distractor objects rather than how small the target is; decoupling object size from clutter density would settle which variable drives the floor.
  • If the floor is largely a crowding effect, then a detection module that first proposes candidate locations could push the O2I limit lower without additional data, since the classifier would no longer need to aggregate over thousands of irrelevant patches.
  • The saliency results already show localization quality degrades faster than classification at small O2I, suggesting that the limit may be even lower for weakly supervised detection than for classification; that gap deserves its own scaling curve.
  • For pathology, the scaling result implies that weakly supervised whole-slide classification at clinically realistic O2I levels would need tens of thousands of slides, so the practical bottleneck is data-efficient inductive biases rather than raw dataset collection.
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

2 major / 5 minor

Summary. The paper introduces two synthetic 'needle-in-a-haystack' testbeds, nMNIST and nCAMELYON, to study image-level binary classification when the object of interest occupies a very small fraction of the image (low object-to-image, O2I, ratio). Using a BagNet-style backbone with controlled receptive fields, capacities, and four global pooling operations, the authors report four empirical observations: (1) there is an O2I limit below which CNNs fail to generalize, and the amount of training data needed for generalization scales rapidly with the inverse of O2I; (2) higher-capacity models generalize better; (3) matching the receptive field to the object size is beneficial on nMNIST but not on nCAMELYON; and (4) global pooling choice affects optimization at the smallest O2I values. The main evidence for the headline scaling law comes from Fig. 6b, which is computed from nMNIST experiments over six training-set sizes and six seeds per configuration.

Significance. The paper addresses a practically important and understudied regime, and the testbed itself is a useful resource: nMNIST provides controlled O2I variation, and the experiments are unusually extensive for a study of this type, with 48 architectures, six seeds per run, and more than 750 reported experiments. If the scaling claim holds, it quantifies a real bottleneck for image-level CNN classifiers on tiny-object tasks. The code-release statement, if honored, would make the testbed reproducible. However, the central quantitative claim about inverse-O2I scaling is currently entangled with a second, uncontrolled factor, distractor count, and the nCAMELYON dataset is too small and too coarsely controlled to provide independent confirmation. The paper's usefulness as a testbed is therefore stronger than the evidence for its headline law.

major comments (2)
  1. [Section 2.1 and Section 3.1 (Fig. 6b)] The nMNIST generator does not vary O2I independently of clutter count. To maintain approximately constant clutter density, the canvas sizes 64, 128, 256, 512, and 1024 pixels are paired with 2, 5, 25, 100, and 400 clutter digits, respectively. Every low-O2I condition is therefore also a high-distractor condition, and the positive decision at 0.075% O2I is 'find one 3 among 401 digits' rather than 'find a 28x28 object in a 1024x1024 image'. The abstract's claim that required training-set size scales rapidly with inverse O2I is read directly from Fig. 6b, which uses only this confounded sweep; no dataset-size sweep is run on nCAMELYON. Please add control experiments that vary O2I at a fixed clutter count (e.g., keep 25 distractors while enlarging the canvas), or vary clutter count at fixed O2I, and report whether the scaling law still holds when the two factors are separated.
  2. [Section 3.1 (Figs. 6, 7, 8)] The reported validation and test accuracies in Figs. 6a, 7, and 8 condition on optimization success: Fig. 6a averages only pipeline configurations that surpassed 70% training accuracy, and Figs. 7 and 8 report only runs that fit the training data, replacing non-fitting runs with random accuracy and a textured bar. This conditioning is stated in the captions, but it means the heatmaps and bar heights reflect generalization conditional on successful training, not the unconditional behavior of the architectures. Please report the fraction of runs that satisfy the conditioning, and show at least one unconditional accuracy result (or state explicitly whether the qualitative conclusions change when all runs are included).
minor comments (5)
  1. [Section 3.1 (Inductive bias paragraph)] The abbreviation 'nMINIST' appears instead of 'nMNIST'; please fix this typo.
  2. [Section 2.1] The sentence 'We maintain approximately constant clutter density over different O2I ratios' would be easier to verify if the actual densities (clutter count multiplied by digit area divided by canvas area) were listed for each O2I value, since the chosen counts give only approximate constancy.
  3. [Figure 6b] Please add explicit axis labels and a description of how 'minimum required training set size' is obtained from the six tested sizes; currently the reader must infer whether the values are interpolated or chosen among the grid points.
  4. [Section 2.1 (nCAMELYON)] The center-wise train/validation/test split is described only in the supplementary material; it should be stated in the main text because it is important for interpreting the nCAMELYON results.
  5. [Tables 1 and 2] There are small typographical issues, e.g., 'differen O2I ratios' in Table 1 and inconsistent use of 'O2IR' versus 'O2I ratio' in the supplementary text; please standardize.

Circularity Check

0 steps flagged · score 1.0 of 10

No circularity: the scaling claims are direct empirical testbed measurements, and the O2I/clutter co-variation is an external-validity confound, not a circular reduction.

full rationale

The paper makes no fitted-parameter-then-predict move and no first-principles derivation that could reduce to its own inputs. The central claim that required training set size grows rapidly as O2I shrinks (Section 3.1, Figures 5a, 6a-b) is an empirical reading of validation and test accuracy over datasets the authors constructed; the O2I levels are realized by canvas resolution and the training-size thresholds are measured, not derived from the definition of O2I. The generation procedure in Section 2.1 does vary clutter count together with canvas size (2, 5, 25, 100, 400 distractors for O2I 19.1% down to 0.075%) to keep clutter density approximately constant; this is a real confound for transferring the headline law to settings where clutter count is fixed, but it is an experimental-design limitation, not a circularity, because neither O2I nor the threshold is defined in terms of the other. Self-citations such as Arpit et al. 2017 and Jastrzebski et al. 2017, which include one of the present authors, appear only as related-work context and are not load-bearing for the O2I claims. The authors themselves flag the main limitation: 'due to the lack of large scale datasets that allow for explicit control of the input signal-to-noise ratios, we were forced to use the synthetically built nMNIST dataset for most of our analysis.' That statement, together with the small nCAMELYON lesion counts, supports treating the paper's results as controlled testbed observations rather than as a derivation whose conclusion is presupposed. No circular step can be exhibited with a specific equation or renamed fitted quantity; therefore the appropriate finding is no significant circularity.

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

The paper does not fit free parameters to data; its conclusions are empirical observations from controlled datasets. The load-bearing assumptions are design choices about what the testbeds represent and what the controlled variable is. The most consequential is the treatment of O2I as isolated from distractor count.

assumptions (4)
  • domain assumption Object-to-image area ratio is an adequate proxy for classification signal-to-noise, with the annotated object as the label-relevant signal.
    The entire experimental design varies O2I by changing canvas or crop size while keeping object size fixed, and interprets lower O2I as harder signal. Section 1 definition and Section 2.1.
  • domain assumption Keeping clutter density constant across O2I levels isolates the effect of object size rather than introducing a confound.
    nMNIST increases clutter count from 2 to 400 as the canvas grows so density stays roughly constant; this changes the number of distractors together with O2I. Section 2.1.
  • domain assumption Positive and negative nCAMELYON crops differ only in the presence of the lesion, not in crop position, tissue composition, or acquisition site.
    Positives are taken as 50 random crops around each contiguous lesion; negatives are random healthy crops filtered for background. Section A.2.
  • domain assumption The chosen pipeline, a BagNet backbone with InstanceNorm, RMSProp, fixed schedule, and global pooling, is representative enough to support conclusions about CNNs generally.
    The paper tests 48 architecture combinations, but all share the same backbone family and optimizer, so conclusions are conditional on that family. Section 2.2 and Section 3.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Needles in Haystacks: On Classifying Tiny Objects in Large Images." pith.science (2026). https://pith.science/paper/RQJ6Y3OP

@misc{pith2026190806037,
  author       = {Pith},
  title        = {Pith review of: Needles in Haystacks: On Classifying Tiny Objects in Large Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RQJ6Y3OP}},
  note         = {Machine review of arXiv:1908.06037}
}
read the original abstract

In some important computer vision domains, such as medical or hyperspectral imaging, we care about the classification of tiny objects in large images. However, most Convolutional Neural Networks (CNNs) for image classification were developed using biased datasets that contain large objects, in mostly central image positions. To assess whether classical CNN architectures work well for tiny object classification we build a comprehensive testbed containing two datasets: one derived from MNIST digits and one from histopathology images. This testbed allows controlled experiments to stress-test CNN architectures with a broad spectrum of signal-to-noise ratios. Our observations indicate that: (1) There exists a limit to signal-to-noise below which CNNs fail to generalize and that this limit is affected by dataset size - more data leading to better performances; however, the amount of training data required for the model to generalize scales rapidly with the inverse of the object-to-image ratio (2) in general, higher capacity models exhibit better generalization; (3) when knowing the approximate object sizes, adapting receptive field is beneficial; and (4) for very small signal-to-noise ratio the choice of global pooling operation affects optimization, whereas for relatively large signal-to-noise values, all tested global pooling operations exhibit similar performance.

Figures

Figures reproduced from arXiv: 1908.06037 by the authors.

Figure 1
Figure 1. Range of Object to Image (O2I) ratios [%] for two medical imaging datasets (CAME￾LYON17 (Ehteshami Bejnordi et al., 2017) and MiniMIAS (Suckling, 1994)) as well as one standard computer vision classification dataset (ImageNet (Deng et al., 2009)). The ratio is defined as O2I = Aobject Aimage , where Aobject and Aimage denote the area of the object and the image, respectively. Together with O2I range, we display exam… view at source ↗
Figure 2
Figure 2. Example images from our nMNIST (a, b) and nCAMELYON (c, d) datasets with different [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Pipeline. Our pipeline is built of three com￾ponents: (1) a CNN extracting topological embedding, (2) a global pooling operation and (3) a binary classifier. See text for details [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (15 more)
Figure 5
Figure 5. Figure 5: Testing the O2I limit. Subfigure (a) depicts the test set performance as a function of training dataset size for the nMNIST dataset, while subfigures (b) and (c) show the test set performance as a function of model capacity for the nMNIST dataset and the nCAMELYON data…
Figure 6
Figure 6. Figure 6: Testing the O2I limit. (a) mean validation set accuracy heatmap for max pooling operation, and (b) minimum required training set size to achieve the noted validation accuracy. We test training set sizes ∈ {1400, 2819, 5638, 7500, 11276, 22552} and report the minimum am…
Figure 7
Figure 7. Figure 7: Inductive bias: for (a) the nMNIST dataset and (b) the nCAMELYON dataset. We report only runs that fit the training data. Otherwise we report random accuracy and depict it with a texture on the bars. (a) nMNIST (b) nCAMELYON [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 10
Figure 10. Figure 10: Example images from our MNIST dataset with different O2I ratios. Top row images [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Example images from our CAMELYON dataset for different crop sizes and O2I ratios. [PITH_FULL_IMAGE:figures/full_fig_p014_11.png]
Figure 12
Figure 12. Figure 12: Impact of the training set balance on model accuracy for different pooling operations and [PITH_FULL_IMAGE:figures/full_fig_p016_12.png]
Figure 13
Figure 13. Figure 13: Impact of the network capacity on the generalization performance dependent on the [PITH_FULL_IMAGE:figures/full_fig_p017_13.png]
Figure 14
Figure 14. Figure 14: Testing the O2I limit. Validation set accuracy heatmap for max, logsumexp, mean and soft attention poolings. We test training set sizes ∈ {1400, 2819, 5638, 7500, 11276, 22552} and report the average validation accuracy. C.4 WEAKLY SUPERVISED OBJECT DETECTION: NMNIST …
Figure 15
Figure 15. Figure 15: Testing the O2I limit. Minimum required training set size to achieve the noted valida￾tion accuracy. We test training set sizes ∈ {1400, 2819, 5638, 7500, 11276, 22552} and report the minimum amount of training examples that achieve a specific validation performance p…
Figure 16
Figure 16. Figure 16: Example images from the nMNIST validation set and their corresponding saliency maps [PITH_FULL_IMAGE:figures/full_fig_p018_16.png]
Figure 17
Figure 17. Figure 17: Average precision for detecting the object of interest using the saliency maps for nMNIST. [PITH_FULL_IMAGE:figures/full_fig_p019_17.png]
Figure 18
Figure 18. Figure 18: Example True Positive Images of nCAMELYON validation sets and their corresponding [PITH_FULL_IMAGE:figures/full_fig_p020_18.png]
Figure 19
Figure 19. Figure 19: Example True Negative Image of nCAMELYON validation sets and corresponding [PITH_FULL_IMAGE:figures/full_fig_p021_19.png]
Figure 20
Figure 20. Figure 20: Example False Negative Image of nCAMELYON validation sets and corresponding [PITH_FULL_IMAGE:figures/full_fig_p021_20.png]
Figure 21
Figure 21. Figure 21: Example False Positive Image of nCAMELYON validation sets and corresponding saliency [PITH_FULL_IMAGE:figures/full_fig_p021_21.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 23 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Dynamic capacity networks

    Amjad Almahairi, Nicolas Ballas, Tim Cooijmans, Yin Zheng, Hugo Larochelle, and Aaron Courville. Dynamic capacity networks. In International Conference on Machine Learning, pp.\ 2549--2558, 2016

    work page 2016

  3. [3]

    BACH: Grand Challenge on Breast Cancer Histology Images

    Guilherme Aresta, Teresa Ara \' u jo, Scotty Kwok, Sai Saketh Chennamsetty, Mohammed Safwan K. P., Alex Varghese, Bahram Marami, Marcel Prastawa, Monica Chan, Michael J. Donovan, Gerardo Fernandez, Jack Zeineh, Matthias Kohl, Christoph Walz, Florian Ludwig, Stefan Braunewell, Maximilian Baust, Quoc Dang Vu, Minh Nguyen Nhat To, Eal Kim, Jin Tae Kwak, Same...

    work page Pith review arXiv 2018

  4. [4]

    Jastrzebski, Nicolas Ballas, David Krueger, Emmanuel Bengio, Maxinder S

    Devansh Arpit, Stanislaw K. Jastrzebski, Nicolas Ballas, David Krueger, Emmanuel Bengio, Maxinder S. Kanwal, Tegan Maharaj, Asja Fischer, Aaron C. Courville, Yoshua Bengio, and Simon Lacoste - Julien. A closer look at memorization in deep networks. In Proceedings of the 34th International Conference on Machine Learning, ICML 2017, Sydney, NSW, Australia, ...

    work page 2017

  5. [5]

    Multiple object recognition with visual attention

    Jimmy Ba, Volodymyr Mnih, and Koray Kavukcuoglu. Multiple object recognition with visual attention. In ICLR , 2015

    work page 2015

  6. [6]

    Benchmarking of image registration methods for differently stained histological slides

    Jiri Borovec, Arrate Munoz-Barrutia, and Jan Kybic. Benchmarking of image registration methods for differently stained histological slides. 10 2018. doi:10.1109/icip.2018.8451040

    arXiv 2018

  7. [7]

    Approximating CNN s with bag-of-local-features models works surprisingly well on imagenet

    Wieland Brendel and Matthias Bethge. Approximating CNN s with bag-of-local-features models works surprisingly well on imagenet. In International Conference on Learning Representations, 2019. URL https://openreview.net/forum?id=SkfMWhAqYQ

    work page 2019

  8. [8]

    Bándi , O

    P. Bándi , O. Geessink , Q. Manson , M. Van Dijk , M. Balkenhol , M. Hermsen , B. Ehteshami Bejnordi , B. Lee , K. Paeng , A. Zhong , Q. Li , F. G. Zanjani , S. Zinger , K. Fukuta , D. Komura , V. Ovtcharov , S. Cheng , S. Zeng , J. Thagaard , A. B. Dahl , H. Lin , H. Chen , L. Jacobsson , M. Hedlund , M. Çetin , E. Halıcı , H. Jackson , R. Chen , F. Both...

    arXiv 2019

Show all 47 references
  1. [9]

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei. ImageNet: A Large-Scale Hierarchical Image Database . In CVPR09, 2009

    2009

  2. [10]

    Babak Ehteshami Bejnordi, Mitko Veta, Paul Johannes van Diest, Bram van Ginneken, Nico Karssemeijer, Geert Litjens, Jeroen A. W. M. van der Laak, , and the CAMELYON16 Consortium. Diagnostic Assessment of Deep Learning Algorithms for Detection of Lymph Node Metastases in Women ...

    2017

  3. [11]

    Deep Learning

    Ian Goodfellow, Yoshua Bengio, and Aaron Courville. Deep Learning. MIT Press, 2016. http://www.deeplearningbook.org

    2016

  4. [12]

    Co-teaching: Robust training of deep neural networks with extremely noisy labels

    Bo Han, Quanming Yao, Xingrui Yu, Gang Niu, Miao Xu, Weihua Hu, Ivor Tsang, and Masashi Sugiyama. Co-teaching: Robust training of deep neural networks with extremely noisy labels. In Advances in Neural Information Processing Systems, pp.\ 8527--8537, 2018

    2018

  5. [13]

    Where's Wally? Walker, 1987

    Martin Handford. Where's Wally? Walker, 1987

    1987

  6. [14]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. CoRR, abs/1512.03385, 2015. URL http://arxiv.org/abs/1512.03385

    2015 arXiv

  7. [15]

    Identity mappings in deep residual networks

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Identity mappings in deep residual networks. In European conference on computer vision, pp.\ 630--645. Springer, 2016

    2016

  8. [16]

    Densely connected convolutional networks

    Gao Huang, Zhuang Liu, Laurens Van Der Maaten, and Kilian Q Weinberger. Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 4700--4708, 2017

    2017

  9. [17]

    Attention-based deep multiple instance learning

    Maximilian Ilse, Jakub M Tomczak, and Max Welling. Attention-based deep multiple instance learning. arXiv preprint arXiv:1802.04712, 2018

    2018 arXiv

  10. [18]

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

    Sergey Ioffe and Christian Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In Proceedings of the 32Nd International Conference on International Conference on Machine Learning - Volume 37, ICML'15, pp.\ 448--456. JMLR.org, ...

    2015

  11. [19]

    Three factors influencing minima in sgd

    Stanis aw Jastrz e bski, Zachary Kenton, Devansh Arpit, Nicolas Ballas, Asja Fischer, Yoshua Bengio, and Amos Storkey. Three factors influencing minima in sgd. arXiv preprint arXiv:1711.04623, 2017

    2017 arXiv

  12. [20]

    Mentornet: Learning data-driven curriculum for very deep neural networks on corrupted labels

    Lu Jiang, Zhengyuan Zhou, Thomas Leung, Li-Jia Li, and Li Fei-Fei. Mentornet: Learning data-driven curriculum for very deep neural networks on corrupted labels. arXiv preprint arXiv:1712.05055, 2017

    2017 arXiv

  13. [21]

    Processing megapixel images with deep attention-sampling models

    Angelos Katharopoulos and Fran c ois Fleuret. Processing megapixel images with deep attention-sampling models. arXiv preprint arXiv:1905.03711, 2019

    1905 arXiv

  14. [22]

    Learning multiple layers of features from tiny images

    Alex Krizhevsky. Learning multiple layers of features from tiny images. Technical report, 2009

    2009

  15. [23]

    Imagenet classification with deep convolutional neural networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. In F. Pereira, C. J. C. Burges, L. Bottou, and K. Q. Weinberger (eds.), Advances in Neural Information Processing Systems 25, pp.\ 1097--1105. Curran Associ...

    2012

  16. [24]

    The handbook of brain theory and neural networks

    Yann LeCun and Yoshua Bengio. The handbook of brain theory and neural networks. chapter Convolutional Networks for Images, Speech, and Time Series, pp.\ 255--258. MIT Press, Cambridge, MA, USA, 1998. ISBN 0-262-51102-9. URL http://dl.acm.org/citation.cfm?id=303568.303704

    1998

  17. [25]

    Gradient-based learning applied to document recognition

    Yann Lecun, Léon Bottou, Yoshua Bengio, and Patrick Haffner. Gradient-based learning applied to document recognition. In Proceedings of the IEEE, pp.\ 2278--2324, 1998

    1998

  18. [26]

    A robust and effective approach towards accurate metastasis detection and pn-stage classification in breast cancer

    Byungjae Lee and Kyunghyun Paeng. A robust and effective approach towards accurate metastasis detection and pn-stage classification in breast cancer. CoRR, abs/1805.12067, 2018. URL http://arxiv.org/abs/1805.12067

    2018 arXiv

  19. [27]

    Microsoft COCO: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll \'a r, and C Lawrence Zitnick. Microsoft COCO: Common objects in context . In European conference on computer vision, pp.\ 740--755. Springer, 2014

    2014

  20. [28]

    Exploring the limits of weakly supervised pretraining

    Dhruv Mahajan, Ross Girshick, Vignesh Ramanathan, Kaiming He, Manohar Paluri, Yixuan Li, Ashwin Bharambe, and Laurens van der Maaten. Exploring the limits of weakly supervised pretraining. In Proceedings of the European Conference on Computer Vision (ECCV), pp.\ 181--196, 2018

    2018

  21. [29]

    Recurrent models of visual attention

    Volodymyr Mnih, Nicolas Heess, Alex Graves, et al. Recurrent models of visual attention. In Advances in neural information processing systems, pp.\ 2204--2212, 2014

    2014

  22. [30]

    Abolafia, Jeffrey Pennington, and Jascha Sohl-Dickstein

    Roman Novak, Yasaman Bahri, Daniel A. Abolafia, Jeffrey Pennington, and Jascha Sohl-Dickstein. Sensitivity and generalization in neural networks: an empirical study. In International Conference on Learning Representations, 2018. URL https://openreview.net/forum?id=HJC2SzZCW

    2018

  23. [31]

    Is object localization for free?-weakly-supervised learning with convolutional neural networks

    Maxime Oquab, L \'e on Bottou, Ivan Laptev, and Josef Sivic. Is object localization for free?-weakly-supervised learning with convolutional neural networks . In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.\ 685--694, 2015

    2015

  24. [32]

    R ^ 2 -cnn: Fast tiny object detection in large-scale remote sensing images

    Jiangmiao Pang, Cong Li, Jianping Shi, Zhihai Xu, and Huajun Feng. R ^ 2 -cnn: Fast tiny object detection in large-scale remote sensing images. arXiv preprint arXiv:1902.06042, 2019

    1902 arXiv

  25. [33]

    Resta , N

    S. Resta , N. Acito , M. Diani , G. Corsini , T. Opsahl , and T. V. Haavardsholm . Detection of small changes in airborne hyperspectral imagery: Experimental results over urban areas. In 2011 6th International Workshop on the Analysis of Multi-temporal Remote Sensing Images (M...

    2011

  26. [34]

    Berens, Cas van den Bogaard, Piergiorgio Cerello, Hao Chen, Qi Dou, Maria Evelina Fantacci, Bram Geurts, Robbert van der Gugten, Pheng Ann Heng, Bart Jansen, Michael M.J

    Arnaud Arindra Adiyoso Setio, Alberto Traverso, Thomas de Bel, Moira S.N. Berens, Cas van den Bogaard, Piergiorgio Cerello, Hao Chen, Qi Dou, Maria Evelina Fantacci, Bram Geurts, Robbert van der Gugten, Pheng Ann Heng, Bart Jansen, Michael M.J. de Kaste, Valentin Kotov, Jack Y...

    2017

  27. [35]

    Simonyan and A

    K. Simonyan and A. Zisserman. Very deep convolutional networks for large-scale image recognition. CoRR, abs/1409.1556, 2014

    2014 arXiv

  28. [36]

    The mammographic image analysis society digital mammogram database''exerpta medica

    John Suckling. The mammographic image analysis society digital mammogram database''exerpta medica. Exerpta Medica International Congress Series, 1069, 01 1994

    1994

  29. [37]

    3d multirater rcnn for multimodal multiclass detection and characterisation of extremely small objects

    Carole H Sudre, Beatriz Gomez Anson, Silvia Ingala, Chris D Lane, Daniel Jimenez, Lukas Haider, Thomas Varsavsky, Lorna Smith, H Rolf J \"a ger, and M Jorge Cardoso. 3d multirater rcnn for multimodal multiclass detection and characterisation of extremely small objects. arXiv p...

    2018 arXiv

  30. [38]

    Revisiting unreasonable effectiveness of data in deep learning era

    Chen Sun, Abhinav Shrivastava, Saurabh Singh, and Abhinav Gupta. Revisiting unreasonable effectiveness of data in deep learning era. In Proceedings of the IEEE international conference on computer vision, pp.\ 843--852, 2017

    2017

  31. [39]

    Tieleman and G

    T. Tieleman and G. Hinton. Lecture 6.5---RmsProp: Divide the gradient by a running average of its recent magnitude . COURSERA: Neural Networks for Machine Learning, 2012

    2012

  32. [40]

    Finding a needle in the haystack: Attention-based classification of high resolution microscopy images

    Naofumi Tomita, Behnaz Abdollahi, Jason Wei, Bing Ren, Arief Suriawinata, and Saeed Hassanpour. Finding a needle in the haystack: Attention-based classification of high resolution microscopy images. arXiv preprint arXiv:1811.08513, 2018

    2018 arXiv

  33. [41]

    Lempitsky

    Dmitry Ulyanov, Andrea Vedaldi, and Victor S. Lempitsky. Instance normalization: The missing ingredient for fast stylization. CoRR, abs/1607.08022, 2016

    2016 arXiv

  34. [42]

    Wei, Laura J

    Jason W. Wei, Laura J. Tafe, Yevgeniy A. Linnik, Louis J. Vaickus, Naofumi Tomita, and Saeed Hassanpour. Pathologist-level classification of histologic patterns on resected lung adenocarcinoma slides with deep neural networks. CoRR, abs/1901.11489, 2019. URL http://arxiv.org/a...

    1901 arXiv

  35. [43]

    Dota: A large-scale dataset for object detection in aerial images

    Gui-Song Xia, Xiang Bai, Jian Ding, Zhen Zhu, Serge Belongie, Jiebo Luo, Mihai Datcu, Marcello Pelillo, and Liangpei Zhang. Dota: A large-scale dataset for object detection in aerial images. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2018

    2018

  36. [44]

    Understanding deep learning requires rethinking generalization

    Chiyuan Zhang, Samy Bengio, Moritz Hardt, Benjamin Recht, and Oriol Vinyals. Understanding deep learning requires rethinking generalization. 2017. URL https://arxiv.org/abs/1611.03530

    2017 arXiv

  37. [45]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  38. [46]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  39. [47]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    1993

Pith tools

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