The (de)biasing Effect of GAN-Based Augmentation Methods on Skin Lesion Images

Majchrowska, Sylwia; Carrasco, Limeros, Sandra

doi:10.1007/978-3-031-16452-1_42

Cited by 7 publications

(8 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An alternative solution is to increase the robustness of the models by artificially introducing appropriately prepared, purpose-biased data into the analyzed dataset 6,19 .…”

Section: Related Work 21 Bias Mitigationmentioning

confidence: 99%

“…Van Simoens and Dhoedt 24 proved that the model, in addition to medically relevant features, was driven by artifacts such as pecular reflections, gel application, and rulers. Mikołajczyk et al 19 showed that there is a strong correlation between artifacts such as black frames, ruler marks, and the skin lesion type (benign/malignant). They showed that models trained on biased data learned spurious correlations resulting in more errors in images with such artifacts.…”

Section: Datasetsmentioning

confidence: 99%

“…The examples of artifacts in this dataset that may introduce bias into the model include hair, frames, rulers, pen marks, or gel drops. Past research showed that frames are correlated with the malignant class and ruler marks with the benign 19 .…”

Section: Datasetsmentioning

confidence: 99%

“…Regarding the skin lesions dataset, the literature shows that black frames are correlated with malignant lesions and ruler marks with the benign class 19 . This finding was additionally supported by authors' research, where additional 2000 samples 3 were 3 All artifact annotations will be publicly available 6/11 hand-labeled.…”

Section: Skin Lesion Classificationmentioning

confidence: 99%

See 3 more Smart Citations

Targeted data augmentation for improving modelrobustness against systematic bias

Mikołajczyk-Bareła,

Ferlin,

Grochowski

2024

Preprint

View full text Add to dashboard Cite

The paper proposes a new and effective bias mitigation method called Targeted Data Augmentation (TDA). Since removingbiases is a tedious, always difficult and, on the other hand, not necessarily an effective approach the authors propose toskillfully insert them, instead. To show the efficiency and to validate the proposed approach, two representative and verydiverse datasets: the dataset of clinical skin lesions and the dataset of male and female faces, were selected to serve as thebenchmarks. The existing biases were first manually examined, identified, and annotated. Then, the use of CounterfactualBias Insertion, has provided the confirmation that the biases like the frame, ruler, and glasses, strongly affect the models. Tomake the models more robust against them, Targeted Data Augmentation was used: in short, the samples were modifiedduring training by randomly inserting biases. The proposed method resulted in a significant decrease in bias measures, morespecifically, from a two-fold to more than 50-fold improvement after training with TDA, with a negligible increase in the error rate.

show abstract

“…An alternative solution is to increase the robustness of the models by artificially introducing appropriately prepared, purpose-biased data into the analyzed dataset 6,19 .…”

Section: Related Work 21 Bias Mitigationmentioning

confidence: 99%

Section: Datasetsmentioning

confidence: 99%

Section: Datasetsmentioning

confidence: 99%

Section: Skin Lesion Classificationmentioning

confidence: 99%

See 2 more Smart Citations

Targeted data augmentation for improving modelrobustness against systematic bias

Mikołajczyk-Bareła,

Ferlin,

Grochowski

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…An example is presented of how different tools for sentiment analysis predict different sentiments for the same utterances but other subject's gender. In one of my very recent studies [75], it was proved that generative models are vulnerable to catching and enhancing biases from data. GANs showed that they not only recreate bias in data but also significantly enhance it.…”

Section: Model Selection and Trainingmentioning

confidence: 99%

Towards Explainable Classifiers Using the Counterfactual Approach - Global Explanations for Discovering Bias in Data

Kwasigroch

2020

Journal of Artificial Intelligence and Soft Computing Research

View full text Add to dashboard Cite

The paper proposes summarized attribution-based post-hoc explanations for the detection and identification of bias in data. A global explanation is proposed, and a step-by-step framework on how to detect and test bias is introduced. Since removing unwanted bias is often a complicated and tremendous task, it is automatically inserted, instead. Then, the bias is evaluated with the proposed counterfactual approach. The obtained results are validated on a sample skin lesion dataset. Using the proposed method, a number of possible bias-causing artifacts are successfully identified and confirmed in dermoscopy images. In particular, it is confirmed that black frames have a strong influence on Convolutional Neural Network’s prediction: 22% of them changed the prediction from benign to malignant.

show abstract