Generalizability vs. Robustness: Adversarial Examples for Medical Imaging

By Magdalini Paschali, Sailesh Conjeti, Fernando Navarro, Nassir Navab. 2018

102 citations recorded by the source corpus.

In this paper, for the first time, we propose an evaluation method for deep learning models that assesses the performance of a model not only in an unseen test scenario, but also in extreme cases of noise, outliers and ambiguous input data. To this end, we utilize adversarial examples, images that fool machine learning models, while looking imperceptibly different from original data, as a measure to evaluate the robustness of a variety of medical imaging models. Through extensive experiments on skin lesion classification and whole brain segmentation with state-of-the-art networks such as Inception and UNet, we show that models that achieve comparable performance regarding generalizability may have significant variations in their perception of the underlying data manifold, leading to an extensive performance gap in their robustness.

Subfields: Artificial Intelligence, Biophysics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Abhijit Guha Roy
2. Christian Wachinger
3. Nassir Navab
4. Ala Al-Fuqaha
5. Junaid Qadir
6. Khansa Rasheed
7. Mohamed Ghaly
8. Adeel Razi

In Biophysics

1. Nassir Navab
2. Carl Vondrick
3. Christian Wachinger
4. Abhijit Guha Roy
5. Adnan Qayyum

Citations in the corpus

• A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI (2019, neutral): “In the medical field (see later section), [6], [43], [63]–[69] have studied methods employing saliency and visual explanations.”
• Explainable, trustworthy, and ethical machine learning for healthcare: A survey (2021, background): “, medical image classification [31], medical image segmentation [70], and as well as using time series medical signals [71].”
• Metric Learning for Adversarial Robustness (2019, methodology): “Following the visualization technique in [28, 30, 33], we investigate what happens to the latent representations as they undergo attack.”
• Data Augmentation with Manifold Exploring Geometric Transformations for Increased Performance and Robustness (2019, methodology): “ G=1.5 G=2 G=2.5 G=3 Fig.4: Examples generated with Random Ane Transformations for Dermot [16] and DDSM [14] for a specic range of Geodesic Distances G. between classes. In previous works, such as [25], attempts have been made to evaluate the robustness of a classier utilizing adversarial examples. ”