English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Robustness of Deep Neural Networks (DNNs) is an important aspect to considerfor their clinical applications. This work examined robustness issue for aDNN-based multi-class classification model via comprehensive experimental andsimulation studies. We constructed a DNN-based multi-class classification modelthat classifies an axial CT image as one of the four body sites of brain & neck(BN), chest (C), abdomen & pelvis (AP), and leg & foot (LF). The model wastrained with whole-body CT images of 37 patients each scanned once and 10 scansof a whole-body phantom with different mAs levels to achieve a F1 score of99.7% averaged over four classes on a testing dataset. We evaluated robustnessof the model against noise perturbations under different mAs levels usingsimulated CT noisy images based on noise power spectrum (NPS) and experimentalCT images acquired in repeated scans of the orange-man phantom. To quantifyrobustness of the trained model, we defined Successful Attack Rate (SAR) as theratio of predictions changed with noises and Confusion Matrix for Robustness(CMR) that represents the percentage of a predicted class without noise beingpredicted to different classes under noise. Besides, we repeatedly trained themodel for 100 times using exactly the same training procedure, hyper-parametersettings, as well as the training and validation dataset to investigaterobustness of models due to inherent randomness in the training process.Finally, to improve model robustness, we employed an adaptive training schemeand demonstrated its effectiveness. Experimental and numerical studiesdemonstrated robustness issue of a DNN-based multi-class classification model.The discoveries highlight the needs for evaluating and ensuring DNN modelrobustness.

Experimental and Numerical Investigations on Robustness of a Deep Neural  Network-based Multi-class Classification Model of CT Images with respect to  Image Noise