FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4286638470> ?p ?o ?g. }

• W4286638470 endingPage "108923" @default.
• W4286638470 startingPage "108923" @default.
• W4286638470 abstract "Convolutional neural networks (CNNs) have been widely applied to medical images. However, medical images are vulnerable to adversarial attacks by perturbations that are undetectable to human experts. This poses significant security risks and challenges to CNN-based applications in clinic practice. In this work, we quantify the scale of adversarial perturbation imperceptible to clinical practitioners and investigate the cause of the vulnerability in CNNs. Specifically, we discover that noise (i.e., irrelevant or corrupted discriminative information) in medical images might be a key contributor to performance deterioration of CNNs against adversarial perturbations, as noisy features are learned unconsciously by CNNs in feature representations and magnified by adversarial perturbations. In response, we propose a novel defense method by embedding sparsity denoising operators in CNNs for improved robustness. Tested with various state-of-the-art attacking methods on two distinct medical image modalities, we demonstrate that the proposed method can successfully defend against those unnoticeable adversarial attacks by retaining as much as over 90% of its original performance. We believe our findings are critical for improving and deploying CNN-based medical applications in real-world scenarios." @default.
• W4286638470 created "2022-07-23" @default.
• W4286638470 creator A5008826782 @default.
• W4286638470 creator A5016047550 @default.
• W4286638470 creator A5017542156 @default.
• W4286638470 creator A5036302697 @default.
• W4286638470 creator A5037675893 @default.
• W4286638470 creator A5047748007 @default.
• W4286638470 creator A5069008709 @default.
• W4286638470 creator A5083081872 @default.
• W4286638470 creator A5085113833 @default.
• W4286638470 creator A5090802305 @default.
• W4286638470 date "2022-12-01" @default.
• W4286638470 modified "2023-10-18" @default.
• W4286638470 title "Robust convolutional neural networks against adversarial attacks on medical images" @default.
• W4286638470 cites W2107901014 @default.
• W4286638470 cites W2117539524 @default.
• W4286638470 cites W2159551006 @default.
• W4286638470 cites W2592929672 @default.
• W4286638470 cites W2618530766 @default.
• W4286638470 cites W2788633781 @default.
• W4286638470 cites W2901158775 @default.
• W4286638470 cites W2919115771 @default.
• W4286638470 cites W2924551358 @default.
• W4286638470 cites W2998175747 @default.
• W4286638470 cites W3019839791 @default.
• W4286638470 cites W3024889886 @default.
• W4286638470 cites W3025404316 @default.
• W4286638470 cites W3087014649 @default.
• W4286638470 cites W3091171660 @default.
• W4286638470 cites W3135970545 @default.
• W4286638470 cites W3174422331 @default.
• W4286638470 doi "https://doi.org/10.1016/j.patcog.2022.108923" @default.
• W4286638470 hasPublicationYear "2022" @default.
• W4286638470 type Work @default.
• W4286638470 citedByCount "6" @default.
• W4286638470 countsByYear W42866384702023 @default.
• W4286638470 crossrefType "journal-article" @default.
• W4286638470 hasAuthorship W4286638470A5008826782 @default.
• W4286638470 hasAuthorship W4286638470A5016047550 @default.
• W4286638470 hasAuthorship W4286638470A5017542156 @default.
• W4286638470 hasAuthorship W4286638470A5036302697 @default.
• W4286638470 hasAuthorship W4286638470A5037675893 @default.
• W4286638470 hasAuthorship W4286638470A5047748007 @default.
• W4286638470 hasAuthorship W4286638470A5069008709 @default.
• W4286638470 hasAuthorship W4286638470A5083081872 @default.
• W4286638470 hasAuthorship W4286638470A5085113833 @default.
• W4286638470 hasAuthorship W4286638470A5090802305 @default.
• W4286638470 hasBestOaLocation W42866384701 @default.
• W4286638470 hasConcept C104317684 @default.
• W4286638470 hasConcept C108583219 @default.
• W4286638470 hasConcept C119857082 @default.
• W4286638470 hasConcept C144024400 @default.
• W4286638470 hasConcept C153180895 @default.
• W4286638470 hasConcept C154945302 @default.
• W4286638470 hasConcept C185592680 @default.
• W4286638470 hasConcept C2779903281 @default.
• W4286638470 hasConcept C36289849 @default.
• W4286638470 hasConcept C37736160 @default.
• W4286638470 hasConcept C38652104 @default.
• W4286638470 hasConcept C41008148 @default.
• W4286638470 hasConcept C41608201 @default.
• W4286638470 hasConcept C55493867 @default.
• W4286638470 hasConcept C63479239 @default.
• W4286638470 hasConcept C81363708 @default.
• W4286638470 hasConcept C95713431 @default.
• W4286638470 hasConcept C97931131 @default.
• W4286638470 hasConceptScore W4286638470C104317684 @default.
• W4286638470 hasConceptScore W4286638470C108583219 @default.
• W4286638470 hasConceptScore W4286638470C119857082 @default.
• W4286638470 hasConceptScore W4286638470C144024400 @default.
• W4286638470 hasConceptScore W4286638470C153180895 @default.
• W4286638470 hasConceptScore W4286638470C154945302 @default.
• W4286638470 hasConceptScore W4286638470C185592680 @default.
• W4286638470 hasConceptScore W4286638470C2779903281 @default.
• W4286638470 hasConceptScore W4286638470C36289849 @default.
• W4286638470 hasConceptScore W4286638470C37736160 @default.
• W4286638470 hasConceptScore W4286638470C38652104 @default.
• W4286638470 hasConceptScore W4286638470C41008148 @default.
• W4286638470 hasConceptScore W4286638470C41608201 @default.
• W4286638470 hasConceptScore W4286638470C55493867 @default.
• W4286638470 hasConceptScore W4286638470C63479239 @default.
• W4286638470 hasConceptScore W4286638470C81363708 @default.
• W4286638470 hasConceptScore W4286638470C95713431 @default.
• W4286638470 hasConceptScore W4286638470C97931131 @default.
• W4286638470 hasFunder F4320332161 @default.
• W4286638470 hasFunder F4320337372 @default.
• W4286638470 hasLocation W42866384701 @default.
• W4286638470 hasOpenAccess W4286638470 @default.
• W4286638470 hasPrimaryLocation W42866384701 @default.
• W4286638470 hasRelatedWork W1561927205 @default.
• W4286638470 hasRelatedWork W2482350142 @default.
• W4286638470 hasRelatedWork W2502115930 @default.
• W4286638470 hasRelatedWork W2965546495 @default.
• W4286638470 hasRelatedWork W3126451824 @default.
• W4286638470 hasRelatedWork W3176240006 @default.
• W4286638470 hasRelatedWork W3191453585 @default.
• W4286638470 hasRelatedWork W4246396837 @default.

• next