Model Fooling Attacks Against Medical Imaging: A Short Survey

Sipola, Tuomo; Puuska, Samir; Kokkonen, Tero

doi:10.11610/isij.4615

Cited by 8 publications

(6 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They changed as few as five pixels in the general CIFAR-10 dataset using a gradient-based dual iterative fusion method [8]. A short survey of model fooling attacks in the medical domain by Sipola et al shows that at least adversarial images and patches have been used in experiments [17]. One-pixel attack is a more advanced method, in which only one pixel of an image is modified in order to fool the classifier [20].…”

Section: Adversarial Attacksmentioning

confidence: 99%

One-Pixel Attack Deceives Computer-Assisted Diagnosis of Cancer

Korpihalkola

Sipola

Puuska

et al. 2021

2021 4th International Conference on Signal Processing and Machine Learning

Self Cite

View full text Add to dashboard Cite

In this article we demonstrate that a state-of-the-art machine learning model predicting whether a whole slide image contains mitosis can be fooled by changing just a single pixel in the input image. Computer vision and machine learning can be used to automate various tasks in cancer diagnostic and detection. If an attacker can manipulate the automated processing, the results can be devastating and in the worst case lead to wrong diagnostic and treatments. In this research one-pixel attack is demonstrated in a real-life scenario with a real tumor dataset. The results indicate that a minor one-pixel modification of a whole slide image under analysis can affect the diagnosis. The attack poses a threat from the cyber security perspective: the one-pixel method can be used as an attack vector by a motivated attacker.

show abstract

Section: Adversarial Attacksmentioning

confidence: 99%

One-Pixel Attack Deceives Computer-Assisted Diagnosis of Cancer

Korpihalkola

Sipola

Puuska

et al. 2021

2021 4th International Conference on Signal Processing and Machine Learning

Self Cite

View full text Add to dashboard Cite

show abstract

“…Furthermore, these threats are real in medical imaging [8], [9]. Following categories of threats have been identified for model fooling against medical imaging: (i) adversarial images, (ii) adversarial patches, (iii) one-pixel attacks and (iv) training process tampering [10]. The first three are in the category of adversarial examples.…”

Section: Introductionmentioning

confidence: 99%

“…They are specifically crafted images that deceive a classifier to make false predictions about input images. If such an attack does not need knowledge of the inside workings of the classifier, it is known as a black-box attack, because the only output needed is the prediction confidence score of the classifier [10], [11]. When an adversarial example is given Fig.…”

Section: Introductionmentioning

confidence: 99%

Color-Optimized One-Pixel Attack Against Digital Pathology Images

Korpihalkola

Sipola

Kokkonen

2021

2021 29th Conference of Open Innovations Association (FRUCT)

Self Cite

View full text Add to dashboard Cite

Modern artificial intelligence based medical imaging tools are vulnerable to model fooling attacks. Automated medical imaging methods are used for supporting the decision making by classifying samples as regular or as having characters of abnormality. One use of such technology is the analysis of whole-slide image tissue samples. Consequently, attacks against artificial intelligence based medical imaging methods may diminish the credibility of modern diagnosis methods and, at worst, may lead to misdiagnosis with improper treatment. This study demonstrates an advanced color-optimized one-pixel attack against medical imaging. A state-of-the-art one-pixel modification is constructed with minimal effect on the pixel's color value. This multi-objective approach mitigates the unnatural coloring of raw none-pixel attacks. Accordingly, it is infeasible or at least cumbersome for a human to see the modification in the image under analysis. This color-optimized one-pixel attack poses an advanced cyber threat against modern medical imaging and shows the importance of data integrity with image analysis.

show abstract

“…Our present analysis is a natural extension to our prior studies related to the attack method. Earlier, we have introduced a list of methods to fool artificial neural networks used in medical imaging [14]. One-pixel attack appeared to be a comprehensive and realistic attack vector, so we decided to further investigate it as a conceptual framework in the medical imaging domain [15].…”

Section: Introductionmentioning

confidence: 99%

Chromatic and spatial analysis of one-pixel attacks against an image classifier

Alatalo,

Korpihalkola,

Sipola

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

One-pixel attack is a curious way of deceiving neural network classifier by changing only one pixel in the input image. The full potential and boundaries of this attack method are not yet fully understood. In this research, the successful and unsuccessful attacks are studied in more detail to illustrate the working mechanisms of a one-pixel attack. The data comes from our earlier studies where we applied the attack against medical imaging. We used a real breast cancer tissue dataset and a real classifier as the attack target. This research presents ways to analyze chromatic and spatial distributions of one-pixel attacks. In addition, we present one-pixel attack confidence maps to illustrate the behavior of the target classifier. We show that the more effective attacks change the color of the pixel more, and that the successful attacks are situated at the center of the images. This kind of analysis is not only useful for understanding the behavior of the attack but also the qualities of the classifying neural network.

show abstract

Model Fooling Attacks Against Medical Imaging: A Short Survey

Cited by 8 publications

References 29 publications

One-Pixel Attack Deceives Computer-Assisted Diagnosis of Cancer

One-Pixel Attack Deceives Computer-Assisted Diagnosis of Cancer

Color-Optimized One-Pixel Attack Against Digital Pathology Images

Chromatic and spatial analysis of one-pixel attacks against an image classifier

Contact Info

Product

Resources

About