Adversarial Attacks on Medical Image Classification.
Simple Summary: As we increasingly rely on advanced imaging for medical diagnosis, it's vital that our computer programs can accurately interpret these images. Even a single mistaken pixel can lead to wrong predictions, potentially causing incorrect medical decisions. This study looks into how these...
| Publicado en: | Cancers Vol. 15; no. 17; pp. 4228 - 4250 |
|---|---|
| Autores principales: | , , |
| Formato: | equations & formulas pictorial research tables/charts Journal Article |
| Publicado: |
MDPI
Sep2023
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=171859966&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 171859966 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 20726694 B74B jtl: Cancers issn: 20726694 maglogo: N pubinfo: dt: Sep2023 vid: 15 iid: 17 pid: 97109 pub: MDPI artinfo: ui: 171859966 171859966 171859966 10.3390/cancers15174228 171859966 ppf: 4228 ppct: 22 formats: tig: atl: Adversarial Attacks on Medical Image Classification. aug: au: Tsai, Min-Jen Lin, Ping-Yi Lee, Ming-En affil: Institute of Information Management, National Yang Ming Chiao Tung University, Hsin-Chu 300, Taiwan sug: subj: Diagnostic Imaging Methods Classification Artificial Intelligence Utilization Technology, Radiologic Computer Environment Human Machine Learning Comparative Studies Descriptive Statistics ab: Simple Summary: As we increasingly rely on advanced imaging for medical diagnosis, it's vital that our computer programs can accurately interpret these images. Even a single mistaken pixel can lead to wrong predictions, potentially causing incorrect medical decisions. This study looks into how these tiny mistakes can trick our advanced algorithms. By changing just one or a few pixels on medical images, we tested how various computer models handled these changes. The findings showed that even small disruptions made it hard for the models to correctly interpret the images. This raises concerns about how reliable our current computer-aided diagnostic tools are and underscores the need for models that can resist such small disturbances. Due to the growing number of medical images being produced by diverse radiological imaging techniques, radiography examinations with computer-aided diagnoses could greatly assist clinical applications. However, an imaging facility with just a one-pixel inaccuracy will lead to the inaccurate prediction of medical images. Misclassification may lead to the wrong clinical decision. This scenario is similar to the adversarial attacks on deep learning models. Therefore, one-pixel and multi-pixel level attacks on a Deep Neural Network (DNN) model trained on various medical image datasets are investigated in this study. Common multiclass and multi-label datasets are examined for one-pixel type attacks. Moreover, different experiments are conducted in order to determine how changing the number of pixels in the image may affect the classification performance and robustness of diverse DNN models. The experimental results show that it was difficult for the medical images to survive the pixel attacks, raising the issue of the accuracy of medical image classification and the importance of the model's ability to resist these attacks for a computer-aided diagnosis. pubtype: Academic Journal doctype: equations & formulas pictorial research tables/charts Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|