A Possible Reason for why Data-Driven Beats Theory-Driven Computer Vision

Alexander Andreopoulos; Iuliia Kotseruba; John K. Tsotsos; Yulong Wu

REVIEW 4 major objections 6 minor 52 references

A Possible Reason for why Data-Driven Beats Theory-Driven Computer Vision

T0 review · 4 major / 6 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 Deep learning's edge over classical vision may come from an unnoticed camera-setting mismatch in benchmark datasets.

desk verdict Useful new sensor-setting statistics for COCO/VOC, but the paper's central claim that classic algorithms lost due to biased operating ranges is under-tested and the conclusion overstates it. read the letter →

arxiv 1908.10933 v2 pith:L3XCRZKZ submitted 2019-08-28 cs.CV

John K. Tsotsos , Iuliia Kotseruba , Alexander Andreopoulos , Yulong Wu This is my paper · ORCID

classification cs.CV

keywords computervisionsensorbiascamerasettingsdatasetdeeplearningtheory-drivenmodelsEXIFmetadataempiricalmethodology

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 argues that the historical dominance of data-driven (deep learning) computer vision over theory-driven methods was not decided on fully fair terms. It claims that widely used benchmarking datasets such as COCO and VOC2007 are overwhelmingly captured with short exposure times and low ISO values, while classical theory-driven algorithms have their own narrow optimal ranges of camera shutter speed and gain that barely overlap with those dataset distributions. Because data-driven models were trained on those same images and the classical algorithms were not matched to their operating ranges, the head-to-head comparisons were unknowingly biased against the theory-driven methods. The paper supports this with controlled performance maps for classic and early learning-based detectors across shutter and gain settings, EXIF statistics of COCO and VOC2007, and evaluations of modern detectors on COCO minival split by sensor settings.

What carries the argument

The central object is the performance surface in the two-dimensional sensor-setting space of shutter speed versus voltage gain, with ISO used as a proxy for gain. Prior controlled studies produced such surfaces for classical detectors, and this paper overlays the empirical distribution of COCO and VOC2007 images onto those surfaces. The overlap comparison carries the argument: tiny overlap means classical methods were tested outside their design envelopes, while deep models were trained inside the dense region of the dataset distribution. The exposure-value formula connecting shutter speed, f-number, and ISO to illumination level is used to bin dataset images by illumination.

What would settle it

Measure the actual overlap between the high-performance regions of the controlled sensor-setting maps and the EXIF-derived distributions of COCO and VOC2007 using direct gain values from cameras rather than ISO proxies, or run the same classical algorithms on COCO and VOC2007 subsets whose EXIF falls inside their measured optimal ranges and check whether their performance rises to the controlled-study levels; if the overlap is not tiny or performance does not improve, the central claim fails.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the sensor settings in commonly used vision datasets strongly favor data-driven models: nearly 90% of COCO images fall into a small cluster of very short exposure times and ISO values up to 800, and VOC2007 shows the same concentration. Controlled experiments with a fixed scene show that classical interest-point, saliency, and object-detection algorithms have orderly, algorithm-specific regions of good performance in the shutter-speed/gain plane, and these regions overlap only slightly with the dataset distributions. Modern deep detectors evaluated on the same COCO images show mean average precision values that fluctuate sharply across sensor-setting bins, indicating a sensitivity that is hard to explain except as a training-distribution effect. The paper concludes that theory-driven algorithms were compared against data-driven algorithms on datasets unrepresentative of the theory-driven algorithms' operating ranges but on which the data-driven algorithms were trained.

Load-bearing premise

The entire overlap argument rests on the assumption that the optimal sensor-setting ranges measured for a handful of interest-point, saliency, and object-detection algorithms transfer to the broad class of theory-driven computer vision models, and that ISO can stand in for the voltage gain values used in the controlled experiments.

Editorial extensions

If this is right

If the claim is right, the widely reported performance gap between deep and classical computer vision methods overstates the true gap for images captured within classical algorithms' optimal operating ranges.
Datasets and evaluation protocols should record and report sensor settings and illumination so that algorithms can be tested inside their design envelopes.
The observed high sensitivity of modern detectors to small changes in shutter speed or ISO contradicts the idea that these models generalize smoothly across imaging conditions.
Head-to-head comparisons that ignore sensor-setting distributions can be systematically biased, and the bias can go unnoticed for years.
A fuller empirical methodology in computer vision would require stating operating ranges for all methods before comparison, similar to product usage instructions.

Reading between the lines

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

One testable extension: constructing a version of COCO or VOC2007 stratified by EXIF so that each theory-driven algorithm is evaluated only on its measured optimal ranges should narrow the gap; a failure to see improvement would weaken the sensor-bias explanation.
The same argument could be applied to other imaging parameters not examined here, such as aperture, focus, white balance, or sensor noise, which may impose similar hidden biases on benchmark comparisons.
The logic also suggests that synthetic robustness evaluations using artificial distortions may be missing the more natural axis of camera-setting variation that real datasets embody.
If the claim generalizes, dataset curation practices could shift from collecting 'in the wild' images to collecting a balanced sweep over sensor settings, changing how both training and evaluation are done.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Colleague,

This paper makes a provocative claim and then partially supports it. The new material — EXIF-derived ISO/exposure distributions for COCO and VOC2007, plus mAP-by-camera-bin for four modern detectors — is a useful addition to the dataset-bias literature. The finding that nearly 90% of COCO images sit at short exposure and ISO 800 or below is concrete and worth knowing. The authors are also right that sensor settings are rarely reported or controlled in benchmarking, and their earlier controlled experiments in [1] and [49] do show that some classic algorithms have narrow operating ranges.

But the central claim, that the AlexNet-era comparisons were 'unknowingly biased' against theory-driven methods, is not established by the evidence here. The paper never runs a classic theory-driven detector on COCO or VOC. The operating-range maps in Figure 1 come from interest-point and saliency detectors (Harris-Affine, Hessian-Affine, MSER, SURF, Scale Saliency). The theory-driven object-recognition systems that actually lost to AlexNet — HOG+SVM, SIFT spatial pyramids, Fisher vectors — are not tested. Section 3 does show DPM and BoW, but the plots are messier and the paper itself says characterization of good-performing settings is more difficult there. Yet the 'tiny overlap' argument in Section 6 is made using Figure 1, not Figure 2. That is a real gap.

The authors also rely on an approximate mapping from voltage gain to ISO ('approximately correspond' in Section 5) without calibration, and EXIF is only available for 59% of COCO and 31% of VOC2007. The mAP heatmaps in Figure 5 have sparse bins with values like 1.00 and 0.00 that are obviously noisy; there are no per-bin counts or error bars. To their credit, the paper acknowledges several of these limitations in Sections 5 and 6. That makes the overclaiming in the Conclusion more frustrating, because the evidence does not support 'were compared on datasets unrepresentative of the theory-driven algorithm operating ranges' as a statement of fact. The honest version would be 'may have been biased' or 'this hypothesis deserves a direct test.'

I agree with the reader's CONDITIONAL verdict. The paper deserves a serious referee because it raises a falsifiable and consequential claim, and the EXIF statistics are reproducible. A good referee should ask for a direct test: run a few representative classic algorithms on a sample of COCO/VOC images (or at least a controlled subset) and compare their performance in the high-overlap vs low-overlap sensor bins, or clearly state that the operating-range transfer is an assumption. Also ask for per-bin sample counts and confidence intervals for the mAP values. With that, the paper could be a solid contribution to methodology.

Who is this for? Anyone working on dataset bias, benchmarking methodology, or the deep-learning takeover narrative. I'd bring it to our reading group and would cite the EXIF statistics if I write about dataset bias. Send it to review: yes, with the expectation of major revision.

Referee Report

4 major / 6 minor

Summary. The paper argues that the historical dominance of deep learning over classical 'theory-driven' computer vision was in part an artifact of biased evaluation methodology. It claims that classical algorithms were tested on datasets whose sensor settings (ISO, exposure time, illumination) lie outside the algorithms' optimal operating ranges, while data-driven models were trained on exactly those distributions. To support this, the paper: (i) reproduces controlled experiments from prior work showing that interest-point, saliency, and some object-detection algorithms have sharply peaked performance as functions of shutter speed and gain; (ii) presents new EXIF statistics from COCO and VOC2007 showing that illumination, ISO, and exposure time distributions are highly uneven, with most images taken under short exposures and low-to-medium ISO; and (iii) evaluates four modern detectors (YOLOv3, Faster R-CNN, Mask R-CNN, RetinaNet) on COCO minival images partitioned by EXIF bins, finding mAP fluctuations across bins. The paper concludes that head-to-head comparisons between theory-driven and data-driven methods were 'unknowingly biased against the theory-driven models.' The central historical claim is stated in the Conclusions (Section 7), while the overlap argument rests on comparing Figure 1 (interest-point detector operating ranges) with Figure 4 (dataset sensor distributions).

Significance. If the historical claim were established, the paper would be a significant contribution to the debate on dataset bias and evaluation methodology in computer vision. The EXIF distribution statistics for COCO and VOC2007 are a novel and useful empirical resource, and the demonstration that modern detectors' performance varies with metadata bins is a cautionary finding that could inform dataset construction and benchmarking. However, the paper's central claim is not directly tested: no classical theory-driven algorithm is evaluated on COCO or VOC images, and the operating-range maps are transferred from a small set of interest-point detectors and four older recognition systems without validation. The paper is honest about some limitations (e.g., the difficulty of comparing across studies and the confounds in Figure 5), but the Conclusions overstate what the evidence supports. The contribution would be strengthened by repositioning the bias claim as a hypothesis with preliminary supporting evidence, or by directly measuring the operating ranges of representative classical competitors on the actual datasets used in landmark comparisons.

major comments (4)

[Section 6 (fourth key point) and Section 7 (Conclusions)] The central claim that theory-driven algorithms in landmark head-to-head comparisons were tested outside their operating ranges is not directly supported by the experiments, because the operating-range measurements in Figures 1 and 2 come from a small set of interest-point/saliency detectors (Harris/Hessian-Affine, MSER, SURF, Scale Saliency) and from DPM, BoW, R-CNN, and SPP-net. None of these are the theory-driven competitors that the paper identifies as having 'begun the surrender' in Section 1 (e.g., HOG+SVM, spatial pyramid matching, Fisher vectors). No classical algorithm is evaluated on COCO or VOC images in this manuscript, so the 'tiny overlap' between good performance ranges and dataset distributions is an extrapolation across both algorithms and tasks.
[Section 5 (ISO proxy and bin edges)] The mapping from the voltage gain used in [1] and [49] to ISO is asserted with the phrase 'approximately correspond' and is never calibrated or validated. ISO on consumer cameras is not a direct measure of sensor gain; it typically combines analog gain, digital gain, and tone-mapping, and its relationship to the gain values on the Bumblebee, FireflyMV, and Flea3 cameras used in the prior studies is not established. Without a validated mapping, the quantitative overlap argument (Section 6) cannot be assessed. Additionally, EXIF data are available for only 59% of COCO and 31% of VOC2007 images (Section 4), and the paper does not analyze whether the missing metadata are missing at random or how selection bias might affect the claimed distributions.
[Section 3 (Figure 2) and Section 6 (overlap claim)] The overlap claim is made by visual inspection rather than by a defined quantitative measure. For the object-detection algorithms, the paper itself states that Figure 2 has 'less orderly structure' and that characterizing good-performing settings is 'more difficult,' yet the conclusion in Section 7 treats these algorithms' operating ranges as known. A quantitative definition of 'good-performing' (e.g., a threshold on mAP or precision-recall), a measured overlap statistic, and confidence intervals would be needed before concluding that the test sets were unrepresentative. The current approach does not allow the reader to distinguish 'tiny overlap' from 'no overlap' or from 'overlap that is small only under an arbitrary threshold.'
[Section 5 (Figure 5) and Section 6 (training bias interpretation)] The evaluation of YOLOv3, Faster R-CNN, Mask R-CNN, and RetinaNet shows mAP fluctuations across ISO/exposure bins, but the paper attributes these fluctuations to training bias without controlling for confounds such as object category composition, image difficulty, and the very small numbers of images in underrepresented bins. The paper acknowledges this ('It is hard to attribute this fluctuation entirely to the sensor bias') but then uses the same results to support the conclusion that data-driven models were favored. A statistical analysis that controls for these factors, or at least reports per-bin image counts and category distributions, is required before the results can bear the weight of the argument.

minor comments (6)

[Section 2 (Figure 1 caption)] The phrase 'all are have different shape' is a grammatical error; it should read 'all have different shapes.'
[Section 4 (Table 1 introduction)] 'not surpisingly' is a typo for 'not surprisingly.'
[Section 6 (final paragraph)] 'no only training, but comparative evaluations, are propery performed' should be 'not only training, but also comparative evaluations, are properly performed.'
[Section 4 (EV formula)] The equation as typeset ('2EV = f 2 t + ISO setting ISO 100') is incomplete and ambiguous; it should be written in standard notation, e.g., EV = log2(f^2/t) - log2(ISO/100).
[Section 5 (Figure 5)] The heatmaps in Figure 5 do not report the number of images per bin; several bins are nearly empty (as visible in Figure 3), so the mAP values in those bins have high variance, and the paper should provide counts or confidence intervals to aid interpretation.
[Abstract and Section 1] The abstract states that 'the tacit change has been the evolution of empirical practice,' but the paper presents a static comparison of current dataset distributions and algorithm sensitivities; it does not measure how empirical practice has changed over time, so the 'change' is assumed rather than demonstrated.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the sensor-bias argument is an empirical comparison that uses independent controlled experiments as inputs, not a derivation that reduces to its own conclusion.

full rationale

The paper's derivation chain is empirical and external: Section 2 reports controlled shutter/gain sweeps from Andreopoulos & Tsotsos [1]; Section 3 reports comparable sweeps from Wu & Tsotsos [49]; Section 4 obtains EXIF-based sensor-setting distributions for COCO and VOC2007; Section 6 directly compares the high-performing regions of Figure 1 with the dataset distributions of Figure 4. None of these steps defines a quantity in terms of the conclusion or fits a parameter and then re-predicts it. The load-bearing prior results are self-cited, but they are peer-reviewed/public controlled experiments with their own datasets and are not fitted to COCO/VOC, so under the stated rules they count as independent evidence rather than circular self-citation. The paper's acknowledged approximations ('we use ISO values as a proxy', 'approximately correspond to the gain values used in the previous works') and its caveats ('this limited test needs to be greatly expanded before finalizing this conclusion'; 'It is difficult to compare our results with the results of the previous works directly') identify external-validity limitations, not circularity. The possible transfer failure from interest-point/saliency detectors to the theory-driven recognition systems in historical head-to-heads is a scientific gap in the empirical argument, not an equation-level reduction. No fabricated or renamed prediction is present; the 'tiny overlap' is a direct comparison of measured regions and measured distributions. Score 0.

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

No new theoretical entities are introduced. The paper's contribution is an empirical argument, so the main ledger entries are the binning choices and the transfer assumptions connecting controlled prior work to public datasets.

free parameters (3)

Illumination level thresholds = low: -4 to 7 EV, mid: 8 to 10 EV, high: 11+ EV
Chosen to 'approximately match' the lighting setup in Wu & Tsotsos [49]; used in Table 1 to conclude that roughly 88-89% of COCO is high/medium illumination. Different thresholds would change the headline percentages.
ISO bin edges = [0, 100, 200, 400, 800, 1600, 3200, 6400, 10000]
Manually selected as an approximate mapping from the 8 gain values in [1] and [49]; used to construct Figures 3 and 5. The mapping is acknowledged as approximate.
Exposure time bins = eight equal intervals on [0,1]s plus a >1s bin
Division of exposure time into 8 bins to mirror the 8 shutter speeds in prior work; binning choices affect which cells in the heatmaps appear as 'good' or 'bad' operating regions.

assumptions (4)

domain assumption ISO sensitivity is a valid proxy for analog gain when comparing sensor settings across different cameras.
Introduced in Section 5 ('we use ISO values as a proxy'); the entire overlap analysis between Figure 1 (gain) and Figure 4 (ISO) depends on this equivalence.
domain assumption EXIF metadata availability is non-informative, so the 59%/58% (COCO) and 31% (VOC2007) subsets are representative.
Only images with EXIF data are analyzed in Section 4; if camera settings correlate with image content or source, the distributions and the mAP-by-bin results are biased.
domain assumption The EV formula plus the assumption of proper exposure estimates scene illumination from EXIF data.
Used in Section 4 to derive illumination levels from f-number, exposure time, and ISO, following Wueller & Fageth [51].
domain assumption The optimal operating ranges measured for five interest-point/saliency algorithms in [1] and four detectors in [49] generalize to theory-driven computer vision models as a class and to the images in COCO/VOC.
Section 6 concludes that 'those algorithms would all perform terribly' based on Figure 1 versus Figure 4, without running those algorithms on COCO/VOC images.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Possible Reason for why Data-Driven Beats Theory-Driven Computer Vision." pith.science (2026). https://pith.science/paper/L3XCRZKZ

@misc{pith2026190810933,
  author       = {Pith},
  title        = {Pith review of: A Possible Reason for why Data-Driven Beats Theory-Driven Computer Vision},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L3XCRZKZ}},
  note         = {Machine review of arXiv:1908.10933}
}

read the original abstract

Why do some continue to wonder about the success and dominance of deep learning methods in computer vision and AI? Is it not enough that these methods provide practical solutions to many problems? Well no, it is not enough, at least for those who feel there should be a science that underpins all of this and that we should have a clear understanding of how this success was achieved. Here, this paper proposes that the dominance we are witnessing would not have been possible by the methods of deep learning alone: the tacit change has been the evolution of empirical practice in computer vision and AI over the past decades. We demonstrate this by examining the distribution of sensor settings in vision datasets and performance of both classic and deep learning algorithms under various camera settings. This reveals a strong mismatch between optimal performance ranges of classical theory-driven algorithms and sensor setting distributions in the common vision datasets, while data-driven models were trained for those datasets. The head-to-head comparisons between data-driven and theory-driven models were therefore unknowingly biased against the theory-driven models.

Figures

Figures reproduced from arXiv: 1908.10933 by the authors.

**Figure 1.** Adapted from Andreopoulos & Tsotsos [1] showing the performance of various descriptors in terms of precision-recall values for each combination of sensor settings (collapsed across all illumination conditions tested). The descriptors tested are (a) Harris-Affine; (b) HessianAffine; (c) MSER; (d) SURF; (e) Scale Saliency. In all plots shutter speed increases from top to bottom and gain increases from left to right.… view at source ↗

**Figure 2.** Results of evaluation of object detection algo [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Distribution of exposure time and ISO in training [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Distributions of exposure times (shutter speeds) [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 45 canonical work pages

[1]

Andreopoulos and J

A. Andreopoulos and J. K. Tsotsos. On sensor bias in exper- imental methods for comparing interest-point, saliency, and recognition algorithms. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence, 34(1), 2011. 2, 3, 5

work page 2011
[49]

Wu and J

Y . Wu and J. Tsotsos. Active control of camera pa- rameters for object detection algorithms. arXiv preprint arXiv:1705.05685, 2017. 3, 4, 5

arXiv 2017
[2]

Andreopoulos and J

A. Andreopoulos and J. K. Tsotsos. 50 years of object recog- nition: Directions forward. Computer vision and image un- derstanding, 117(8), 2013. 1

work page 2013
[3]

Azulay and Y

A. Azulay and Y . Weiss. Why do deep convolutional net- works generalize so poorly to small image transformations? arXiv preprint arXiv:1805.12177, 2018. 6

arXiv 2018
[4]

D. H. Ballard and C. M. Brown. Computer Vision. Prentice- Hall, 1982. 1

work page 1982
[5]

H. Bay, T. Tuytelaars, and L. Van Gool. Surf speeded up robust features. In Proceedings of the European Conference on Computer Vision (ECCV), 2006. 2

work page 2006
[6]

N. D. Bruce and J. K. Tsotsos. Saliency, attention, and vi- sual search: An information theoretic approach. Journal of vision, 9(3), 2009. 3

work page 2009
[7]

Z. Che, A. Borji, G. Zhai, X. Min, G. Guo, and P. Le Callet. GazeGAN: A Generative Adversarial Saliency Model based on Invariance Analysis of Human Gaze During Scene Free Viewing. arXiv preprint arXiv:1905.06803, 2019. 6

arXiv 1905

Show all 52 references

[8]

Dodge and L

S. Dodge and L. Karam. A study and comparison of human and deep learning recognition performance under visual dis- tortions. In In Proceedings of the 26th International Confer- ence on Computer Communication and Networks (ICCCN) ,
[9]

Everingham, L

M. Everingham, L. Van Gool, C. K. I. Williams, J. Winn, and A. Zisserman. The PASCAL Visual Object Classes Challenge 2007 (VOC2007) Results. http://www.pascal- network.org/challenges/VOC/voc2007/workshop/index.html. 4

2007
[10]

Everingham, A

M. Everingham, A. Zisserman, C. K. Williams, L. Van Gool, M. Allan, C. M. Bishop, O. Chapelle, N. Dalal, T. Dese- laers, G. Dork ´o, et al. The 2005 PASCAL Visual Object Classes Challenge. In Machine Learning Challenges Work- shop, 2005. 2

2005
[11]

Faugeras

O. Faugeras. Three-Dimensional Computer Vision: A Geo- metric Viewpoint. MIT press, 1993. 1

1993
[12]

P. F. Felzenszwalb, R. B. Girshick, D. McAllester, and D. Ra- manan. Object detection with discriminatively trained part- based models. IEEE Transactions on Pattern Analysis and Machine Intelligence, 32(9), 2009. 3

2009
[13]

Geirhos, C

R. Geirhos, C. R. Temme, J. Rauber, H. H. Sch ¨utt, M. Bethge, and F. A. Wichmann. Generalisation in humans and deep neural networks. In Advances in Neural Informa- tion Processing Systems (NeuroIPS), 2018. 6

2018
[14]

Girshick, J

R. Girshick, J. Donahue, T. Darrell, and J. Malik. Rich fea- ture hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014. 3

2014
[15]

K. He, G. Gkioxari, P. Doll ´ar, and R. Girshick. Mask R- CNN. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017. 5

2017
[16]

K. He, X. Zhang, S. Ren, and J. Sun. Spatial Pyramid Pool- ing in Deep Convolutional Networks for Visual Recognition. IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 37(9), 2015. 3

2015
[17]

G. E. Hinton, C. K. Williams, and M. D. Revow. Adaptive elastic models for hand-printed character recognition. In Ad- vances in Neural Information Processing Systems, 1992. 1

1992
[18]

B. K. P. Horn. Robot vision. MIT press, 1986. 1

1986
[19]

L. Itti, C. Koch, and E. Niebur. A model of saliency-based visual attention for rapid scene analysis. IEEE Transactions on Pattern Analysis & Machine Intelligence, (11), 1998. 3

1998
[20]

Kadir and M

T. Kadir and M. Brady. Saliency, scale and image descrip- tion. International Journal of Computer Vision, 45(2), 2001. 2

2001
[21]

J. J. Koenderink. Solid shape. MIT press, 1990. 1

1990
[22]

Krizhevsky, I

A. Krizhevsky, I. Sutskever, and G. E. Hinton. Imagenet classiﬁcation with deep convolutional neural networks. In Advances in Neural Information Processing Systems , 2012. 2

2012
[23]

Y . LeCun. The MNIST Database of Handwritten Digits. http://yann.lecun.com/exdb/mnist/, 1998. 1

1998
[24]

T.-Y . Lin, P. Goyal, R. Girshick, K. He, and P. Doll´ar. Focal Loss for Dense Object Detection. InProceedings of the IEEE International Conference on Computer Vision (ICCV), 2017. 4, 5

2017
[25]

T.-Y . Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ra- manan, P. Doll´ar, and C. L. Zitnick. Microsoft COCO: Com- mon Objects in Context. In Proceedings of the European Conference on Computer Vision (ECCV), 2014. 4

2014
[26]

W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C.-Y . Fu, and A. C. Berg. SSD: Single Shot Multibox Detector. In Proceedings of the European Conference on Computer Vi- sion (ECCV), 2016. 4

2016
[27]

D. Marr. Vision: A Computational Investigation into the Hu- man Representation and Processing of Visual Information . MIT Press, 1982. 1

1982
[28]

Matas, O

J. Matas, O. Chum, M. Urban, and T. Pajdla. Robust Wide- Baseline Stereo from Maximally Stable Extremal Regions. Image and Vision Computing, 22(10), 2004. 2

2004
[29]

Mikolajczyk and C

K. Mikolajczyk and C. Schmid. Scale & afﬁne invariant in- terest point detectors. International Journal of Computer Vi- sion, 60(1), 2004. 2

2004
[30]

Nickolls and W

J. Nickolls and W. J. Dally. The GPU Computing Era. IEEE Micro, 30(2), 2010. 1

2010
[31]

Redmon and A

J. Redmon and A. Farhadi. YOLOv3: An Incremental Im- provement. arXiv preprint arXiv:1804.02767, 2018. 4, 5

2018 arXiv
[32]

S. Ren, K. He, R. Girshick, and J. Sun. Faster r-cnn: Towards real-time object detection with region proposal networks. In Advances in Neural Information Processing Systems , 2015. 5

2015
[33]

Rodner, M

E. Rodner, M. Simon, R. B. Fisher, and J. Denzler. Fine- grained recognition in the noisy wild: Sensitivity analysis of convolutional neural networks approaches. arXiv preprint arXiv:1610.06756, 2016. 6

2016 arXiv
[34]

Rosenfeld

A. Rosenfeld. Digital Picture Processing. Academic press,
[35]

Rosenfeld, R

A. Rosenfeld, R. Zemel, and J. K. Tsotsos. The elephant in the room. arXiv preprint arXiv:1808.03305, 2018. 6

2018 arXiv
[36]

P. Roy, S. Ghosh, S. Bhattacharya, and U. Pal. Effects of degradations on deep neural network architectures. arXiv preprint arXiv:1807.10108, 2018. 6

2018 arXiv
[37]

Russakovsky, J

O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, S. Ma, Z. Huang, A. Karpathy, A. Khosla, M. Bernstein, et al. ImageNet Large Scale Visual Recognition Challenge. International Journal of Computer Vision, 115(3), 2015. 2

2015
[38]

G. Salton. Introduction to Modern Information Retrieval . McGraw-Hill College, 1983. 3

1983
[39]

Shapiro, editor

S. Shapiro, editor. The Encyclopedia of Artiﬁcial Intelli- gence. John Wiley & Sons, New York, 2 edition, 1992. 1

1992
[40]

Sirovich and M

L. Sirovich and M. Kirby. Low-dimensional procedure for the characterization of human faces. Journal of Optical So- ciety of America, 4(3), 1987. 1

1987
[41]

J. Su, D. V . Vargas, and K. Sakurai. One pixel attack for fooling deep neural networks. IEEE Transactions on Evolu- tionary Computation, 2019. 6

2019
[42]

Szeliski

R. Szeliski. Computer Vision: Algorithms and Applications. Springer, 2010. 1

2010
[43]

Tommasi, N

T. Tommasi, N. Patricia, B. Caputo, and T. Tuytelaars. A deeper look at dataset bias. In Domain Adaptation in Com- puter Vision Applications. Springer, 2017. 4

2017
[44]

A. W. Tow, S. Shirazi, J. Leitner, N. S¨underhauf, M. Milford, and B. Upcroft. A robustness analysis of deep q networks. In Proceedings of the Australasian Conference on Robotics and Automation, 2016. 6

2016
[45]

J. K. Tsotsos. A complexity level analysis of vision. Inter- national Journal of Computer Vision, 1, 1987. 1

1987
[46]

J. K. Tsotsos. The complexity of perceptual search tasks. In Proceedings of the International Joint Conference on Artiﬁ- cial Intelligence (IJCAI), volume 89, 1989. 1

1989
[47]

M. A. Turk and A. P. Pentland. Face Recognition Using Eigenfaces. In Proceedings of the IEEE Conference on Com- puter Vision and Pattern Recognition (CVPR), 1991. 1

1991
[48]

J. R. Uijlings, K. E. Van De Sande, T. Gevers, and A. W. Smeulders. Selective search for object recognition. Interna- tional Journal of Computer Vision, 104(2), 2013. 3

2013
[50]

Wueller and R

D. Wueller and R. Fageth. Statistic analysis of millions of digital photos. In Digital Photography IV , volume 6817,
[51]

Wueller and R

D. Wueller and R. Fageth. Statistic analysis of millions of digital photos 2017. Electronic Imaging, 2018(5), 2018. 4

2017
[52]

S. Zucker. Vision, early. In S. Shapiro, editor, The Ency- clopedia of Artiﬁcial Intelligence. John Wiley & Sons, New York, 2 edition, 1987. 1

1987

Pith tools

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

[1] [1]

Andreopoulos and J

A. Andreopoulos and J. K. Tsotsos. On sensor bias in exper- imental methods for comparing interest-point, saliency, and recognition algorithms. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence, 34(1), 2011. 2, 3, 5

work page 2011

[2] [49]

Wu and J

Y . Wu and J. Tsotsos. Active control of camera pa- rameters for object detection algorithms. arXiv preprint arXiv:1705.05685, 2017. 3, 4, 5

arXiv 2017

[3] [2]

Andreopoulos and J

A. Andreopoulos and J. K. Tsotsos. 50 years of object recog- nition: Directions forward. Computer vision and image un- derstanding, 117(8), 2013. 1

work page 2013

[4] [3]

Azulay and Y

A. Azulay and Y . Weiss. Why do deep convolutional net- works generalize so poorly to small image transformations? arXiv preprint arXiv:1805.12177, 2018. 6

arXiv 2018

[5] [4]

D. H. Ballard and C. M. Brown. Computer Vision. Prentice- Hall, 1982. 1

work page 1982

[6] [5]

H. Bay, T. Tuytelaars, and L. Van Gool. Surf speeded up robust features. In Proceedings of the European Conference on Computer Vision (ECCV), 2006. 2

work page 2006

[7] [6]

N. D. Bruce and J. K. Tsotsos. Saliency, attention, and vi- sual search: An information theoretic approach. Journal of vision, 9(3), 2009. 3

work page 2009

[8] [7]

Z. Che, A. Borji, G. Zhai, X. Min, G. Guo, and P. Le Callet. GazeGAN: A Generative Adversarial Saliency Model based on Invariance Analysis of Human Gaze During Scene Free Viewing. arXiv preprint arXiv:1905.06803, 2019. 6

arXiv 1905