A survey on Image Data Augmentation for Deep Learning

Connor Shorten, Taghi M. Khoshgoftaar · Journal Of Big Data · 2019 · machine-learning · 10.1186/s40537-019-0197-0 · record shorten2019-a-survey-on-image-data-augmentatio
tag layer · index · concepts · authors · עברית

Where this record's text came from. 2026-08-20 and 2026-08-26 schema-0.3 passes: every quote verified as an exact substring of the staged text and anchored at merge time; concept terms verified present; every edge endpoint an existing node.

This record’s own method is review, and the tags below follow from that. A review’s findings, quantities and statistics are things stated IN this paper, and some of them report results obtained elsewhere. No tag on this page names a different author. Measured across every review in this corpus, 81% of a review’s tags carry no citation at all — most of them are the review speaking. The corpus does not adjudicate the rest: read an unmarked tag as a claim this paper makes, not necessarily one it made first.

How to read this page. The article itself is not reproduced here — this paper is not cleared for redistribution, so what follows is the tag layer alone. The evidence sentences are withheld with the article. The table at the foot lists every tag the record carries; each row names the facet it belongs to, and the whose words column says whether the tag records the paper's own wording or the tagger's sentence about the paper. Every tag links to each other paper in the corpus carrying it. The machine-readable edition of the same thing is the TEI file beside this one, where the tags live in standOff, and the article is at its publisher via the link below. When we have changed our mind. Some entries in the table begin with a word in capitals and the reason for it. EVIDENCE CORRECTED — the finding still stands, but we had quoted the wrong sentence from the paper, and the quotation has been replaced with the right one. RETRACTED — the entry itself was wrong; it stays on the page, marked, with the reason, rather than being deleted. WITHDRAWN AND REPLACED — the entry was wrong, and a corrected version of it appears in the entries below. Each mark carries its own reason, in brackets, because the reasons differ from one entry to the next. An entry we have WITHDRAWN is removed from this page and kept in the record, with its reason. A line reading Audit and a date is a later re-reading of the entry against its whole sentence: DOUBTFUL means it was kept, and the reason says what is loose about it. Quotation marks inside the tagger's sentence. Where a tag is the tagger's own sentence, a phrase in quotation marks inside it is borrowed from the paper — the tagger's wording around it is not. The paper's own sentence is the evidence quote for that row; on this page it is withheld with the article, and the article is at its publisher. Nothing inside quotation marks anywhere on this page is ours: if it is quoted, it is the paper's. Whose words a concept carries. A concept entry records a term the paper itself uses, so the term, the sense given for it, and its evidence are the paper's own words, quoted where the paper fixes them. This is the opposite of a sense entry in the tag table, which is the tagger's sentence about how the paper uses a term. Where a concept entry could not be grounded in the paper's words, the entry says so on its own row.

Other fields define these differently the same word, defined another way in another discipline
Transfer Learning 6 fields · overfitting 6 fields · Data Augmentation 2 fields

Tags on this paper

Every tag the record carries, each linking to every other paper in the corpus that carries it. A tag appears here whether or not its words occur in the article — the spans below are only the ones a sentence in this paper licenses. Of the 25 concept tags on this paper, all 25 are grounded in a quoted sentence. 11 carry the tagger's gloss of what the term means in this paper, which is not something the article can license.

Data Augmentation Generalizability overfitting data warping oversampling Dropout Batch normalization Transfer Learning Pretraining translational invariances training dataset regularization technique big data GAN samples safety class imbalance Random Oversampling SMOTE refined labels Neural Style Transfer positional bias kernel filters PatchShuffle color space transformations noise injection AUGMENTATION IS AN ASSUMPTION, AND THE PAPER STATES IT: that more information can be extracted from the dataset already held. Everything else follows from whether that is true of a given dataset WHAT AUGMENTATION IS ACTUALLY TEACHING: the invariances a recogniser must survive - viewpoint, lighting, occlusion, background, scale - are baked into the DATA rather than into the architecture SIMULATING BIG DATA RATHER THAN OBTAINING IT: overfitting is not much of a problem when data is abundant, so augmentation gives a limited dataset the characteristics of an abundant one rotation degree review

Method, as described: a survey organised by where the augmentation intervenes, with medical imaging as the motivating limited-data case

Legend

claim type	`claim_type`	the shape of the claim: existence, mechanism, method, critique, prediction, null
concept edge	`concept_edges`	an asserted relation between two concepts, with the sentence asserting it
concept	`concepts`	an idea the paper works with, quoted where the paper uses it
content tag	`content_tags`	a finding, definition, quantity, statistic or variable named in the paper
contribution	`contribution`	what the paper claims to add
limitation	`limitations`	what the paper says its evidence cannot support
mechanism	`mechanisms`	how the reported effect is said to come about
method	`method`	how the paper produced its evidence
object of study	`object_of_study`	what the paper is about — the thing it investigates
phenomenon	`phenomenon`	the effect or regularity under study
research question	`research_question`	the question the paper sets itself
sense	`senses`	a distinct use of a term, quoted where the paper fixes it
transfer	`transfers`	a method or concept carried in from another field

The paper

The full text is not reproduced here. This article is not cleared for redistribution, so the page carries the tag layer only. Its evidence sentences are withheld; the grounding status remains with each tag, and the article itself is at https://doi.org/10.1186/s40537-019-0197-0. The reading view exists for the papers whose licence permits it.

Every tag on this page

claim_type 1 · concept_edges 5 · concepts 25 · content_tags 13 · contribution 1 · limitations 12 · mechanisms 10 · method 1 · object_of_study 1 · phenomenon 1 · research_question 1 · senses 3 · transfers 2

Relations on this page carry a quotation for the claim. Where a relation was separately evidenced — and so tested against the vocabulary for its type — the row says so; 1 of 5 relations here carries that mark. The rest are the tagger’s reading of the sentence quoted beside them.

This record holds no relations between claims or controls. A layer is empty when the tagger recorded nothing in it — which is not the same as the paper having nothing to record.

facet	tag	field	conf.	whose words
object_of_study	image Data Augmentation techniques, from geometric transforms to GANs, as a remedy for overfitting on limited data Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`object_of_study`		the paper's words
research_question	what to do about overfitting when the application domain has no big data to be had Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`research_question`		the paper's words
contribution	a taxonomy of image augmentation split into data warping and oversampling, together with the statement that the split is not exclusive Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`contribution`		the paper's words
claim_type	method Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`claim_type`		the paper's words
phenomenon	overfitting - a model learning a function of very high variance that models its training data perfectly Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`phenomenon`		the paper's words
method	review Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: review	`method`		the paper's words
mechanisms	mechanisms Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: AUGMENTATION IS AN ASSUMPTION, AND THE PAPER STATES IT: that more information can be extracted from the dataset already held. Everything else follows from whether that is true of a given dataset	`mechanisms[0]`	0.95	the tagger's words
mechanisms	mechanisms Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: WHAT AUGMENTATION IS ACTUALLY TEACHING: the invariances a recogniser must survive - viewpoint, lighting, occlusion, background, scale - are baked into the DATA rather than into the architecture	`mechanisms[1]`	0.95	the tagger's words
mechanisms	mechanisms Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: SIMULATING BIG DATA RATHER THAN OBTAINING IT: overfitting is not much of a problem when data is abundant, so augmentation gives a limited dataset the characteristics of an abundant one	`mechanisms[2]`	0.95	the tagger's words
mechanisms	an unsafe transformation could be made useful rather than harmful by softening the label, so the model learns uncertainty instead of a falsehood - but building those labels is too costly Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`mechanisms[3]`	0.95	the tagger's words
mechanisms	oversampling works by re-balancing the class distribution so the model is not biased toward answering with the majority class Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`mechanisms[4]`	0.9	the tagger's words
mechanisms	medical imaging is short of data FOR STRUCTURAL REASONS, and the paper names rarity first: `Quotation withheld`. The remaining three reasons it lists - patient privacy, the need for expert labellers, and the cost of the scans - are separated from this clause by a journal running header in the staged text and are not inside this span. Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`mechanisms[5]`	0.95	the tagger's words
mechanisms	rotation safety is set by one parameter: small rotations help digit recognition and large ones destroy the label, so the same transform is safe and unsafe at different magnitudes Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`mechanisms[6]`	0.95	the tagger's words
mechanisms	cropping and translation differ in what they preserve - cropping shrinks the input, translation keeps the spatial dimensions and pads the vacated space Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`mechanisms[7]`	0.95	the tagger's words
mechanisms	colour transformations trade colour bias for spatial characteristics, which is a gain only where colour is not the distinguishing feature Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`mechanisms[8]`	0.95	the tagger's words
mechanisms	a kernel filter can be rebuilt as a convolution layer with the right padding and activation, so the augmentation duplicates what the network already does Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`mechanisms[9]`	0.95	the tagger's words
limitations	AUGMENTATION CANNOT CREATE A CLASS THAT IS ABSENT: no method surveyed will produce a golden retriever from a dataset of bulldogs. It can lessen biases of lighting, occlusion, scale and background, and it cannot supply missing content Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[0]`	0.95	the tagger's words
limitations	The stated bias limit in the paper's own words, which bounds every technique it surveys Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[1]`	0.85	the tagger's words
limitations	The survey covers IMAGE data only, with transfer to other data domains asserted rather than shown Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[2]`	0.95	the tagger's words
limitations	The taxonomy the survey is organised by does not partition its own subject - warping and oversampling are stated not to be mutually exclusive Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[3]`	0.95	the tagger's words
limitations	safety is domain-dependent, which is precisely what makes a generalisable augmentation policy hard Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[4]`	0.95	the tagger's words
limitations	there is NO image processing function that cannot flip a label at some distortion magnitude - the strongest form of the claim, stated by the authors Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[5]`	0.95	the tagger's words
limitations	constructing refined labels for every unsafe augmentation is computationally expensive, which is why the cheap option is to avoid unsafe transforms Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[6]`	0.9	the tagger's words
limitations	many experiments deliberately restrict themselves to a subset of a large dataset in order to SIMULATE a limited-data problem rather than face a real one Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[7]`	0.9	the tagger's words
limitations	kernel filters are very close to the internal mechanism of a CNN, which is the argument for making them a layer rather than a dataset operation Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[8]`	0.95	the tagger's words
limitations	in medical imaging the gap between training and test data is more complex than position and translation, so geometric transformations have limited scope there Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[9]`	0.95	the tagger's words
limitations	colour transformations can discard the information the task depends on - darkening may hide the object entirely Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[10]`	0.95	the tagger's words
limitations	transformations that may alter the label, such as translation or random cropping, must be checked by hand Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`limitations[11]`	0.9	the tagger's words
transfers	SMOTE and Borderline-SMOTE were built to interpolate new points for imbalanced TABULAR data and are carried over to images Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`transfers[0]`	0.95	the tagger's words
transfers	transfer learning is explained by the anecdote that a pianist learns guitar faster, and augmentation by analogy with imagination and dreaming Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`transfers[1]`	0.9	the tagger's words
concepts	Data Augmentation Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: Data Augmentation	`concepts[0]`	0.95	the paper's words
concepts	Generalizability Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: Generalizability	`concepts[1]`	0.95	the paper's words
concepts	overfitting Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: overfitting	`concepts[2]`	0.95	the paper's words
concepts	data warping Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: data warping	`concepts[3]`	0.95	the paper's words
concepts	oversampling Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: oversampling	`concepts[4]`	0.95	the paper's words
concepts	Dropout Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: Dropout	`concepts[5]`	0.95	the paper's words
concepts	Batch normalization Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: Batch normalization	`concepts[6]`	0.95	the paper's words
concepts	Transfer Learning Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: Transfer Learning	`concepts[7]`	0.95	the paper's words
concepts	Pretraining Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: Pretraining	`concepts[8]`	0.85	the paper's words
concepts	translational invariances Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: translational invariances	`concepts[9]`	0.95	the paper's words
concepts	training dataset Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: training dataset	`concepts[10]`	0.95	the paper's words
concepts	regularization technique Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: regularization technique	`concepts[11]`	0.95	the paper's words
concepts	big data Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: big data	`concepts[12]`	0.95	the paper's words
concepts	GAN samples Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: GAN samples	`concepts[13]`	0.95	the paper's words
concepts	safety Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: safety	`concepts[14]`	0.95	the paper's words
concepts	class imbalance Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: class imbalance	`concepts[15]`	0.95	the paper's words
concepts	Random Oversampling Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: Random Oversampling	`concepts[16]`	0.9	the paper's words
concepts	SMOTE Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: SMOTE	`concepts[17]`	0.95	the paper's words
concepts	refined labels Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: refined labels	`concepts[18]`	0.9	the paper's words
concepts	Neural Style Transfer Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: Neural Style Transfer	`concepts[19]`	0.9	the paper's words
concepts	positional bias Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: positional bias	`concepts[20]`	0.95	the paper's words
concepts	kernel filters Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: kernel filters	`concepts[21]`	0.95	the paper's words
concepts	PatchShuffle Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: PatchShuffle	`concepts[22]`	0.9	the paper's words
concepts	color space transformations Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: color space transformations	`concepts[23]`	0.95	the paper's words
concepts	noise injection Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: noise injection	`concepts[24]`	0.9	the paper's words
senses	Data Augmentation - A DATA-SPACE SOLUTION, defined by WHERE it intervenes rather than by what it does: every other remedy for overfitting changes the model, and this one changes the dataset Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`senses[0]`	0.95	the tagger's words
senses	data warping - the label is PRESERVED and an existing image is transformed Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`senses[1]`	0.95	the tagger's words
senses	oversampling - a NEW instance is synthesised and added. The paper draws the distinction and then says in the next breath that it does not partition: GAN samples can be stacked with random cropping Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`senses[2]`	0.95	the tagger's words
concept_edges	data warping —distinct-from→ oversampling Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. holds when: the two are NOT mutually exclusive - GAN samples can be stacked with random cropping, so a single augmentation can be both · strength of the relation: 0.8 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says the two augmentations do NOT form a mutually exclusive dichotomy; distinct-from is hedged. also in: data warping · oversampling	`concept_edges[0]`		the tagger's words
concept_edges	Transfer Learning —distinct-from→ Pretraining Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. strength of the relation: 1 Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but endpoint by anaphora: 'This differs from Transfer Learning' - 'This' = Pretraining (previous sentences). also in: Transfer Learning · Pretraining	`concept_edges[1]`		the tagger's words
concept_edges	Dropout —is-a→ regularization technique Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. strength of the relation: 1 also in: Dropout · regularization technique	`concept_edges[2]`		the tagger's words
concept_edges	Batch normalization —is-a→ regularization technique Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. strength of the relation: 1 also in: Batch normalization · regularization technique	`concept_edges[3]`		the tagger's words
concept_edges	Data Augmentation —causes→ overfitting Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. direction on the measure: reduces · the relation itself is separately evidenced · strength of the relation: 1 · direction on the measure (stated by the tagger, no quotation): reduces · the relation stated: augmentation gives a limited dataset the characteristics of big data, and overfitting is not much of an issue when data is abundant also in: Data Augmentation · overfitting	`concept_edges[4]`		the tagger's words
content_tags	AlexNet's experiments used augmentation to enlarge the dataset by a factor of 2048. Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.findings[0]`	0.95	the tagger's words
content_tags	GAN-based augmentation beat classic augmentation on liver lesion classification on both sensitivity and specificity Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.findings[1]`	0.95	the tagger's words
content_tags	Horizontal flipping is much more common than vertical flipping. Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.findings[2]`	0.9	the tagger's words
content_tags	converting to grayscale costs about three points of classification accuracy against RGB on ImageNet and PASCAL VOC Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.findings[3]`	0.95	the tagger's words
content_tags	in a head-to-head comparison of geometric against photometric augmentation, cropping produced the most accurate classifier Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.findings[4]`	0.95	the tagger's words
content_tags	randomly swapping pixels inside a small window lowered CIFAR-10 error against the same architecture without it Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.findings[5]`	0.95	the tagger's words
content_tags	Blurring training images could make models more resistant to motion blur at test time. Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.findings[6]`	0.9	the tagger's words
content_tags	the safety of an augmentation defined as label preservation, with the cat/dog against 6/9 example Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: safety	`content_tags.definitions[0]`	0.95	the tagger's words
content_tags	the factor by which AlexNet's augmentation multiplied the dataset, and the error reduction claimed for it Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.quantities[0]`	0.95	the tagger's words
content_tags	sensitivity and specificity before and after GAN-based augmentation on liver lesion classification Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.quantities[1]`	0.95	the tagger's words
content_tags	the CIFAR-10 error rate with and without PatchShuffle regularization, and the hyperparameters that achieved it Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.quantities[2]`	0.95	the tagger's words
content_tags	the accuracy drop from grayscale against RGB Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above.	`content_tags.quantities[3]`	0.95	the tagger's words
content_tags	the rotation degree parameter, on which the safety of a rotation augmentation depends Quotation withheld — this record is not cleared for redistribution. The sentence is in the paper at the link above. also in: rotation degree	`content_tags.variables[0]`	0.95	the tagger's words

Interpretive choices

Tags that could not be placed in the text:

none — every tag was located in the text

What was deliberately not tagged, and why, is recorded per paper by the tagger. Where a paper has no such record, this page says so rather than implying that nothing was declined.