More Like this

1. A Study of Adversarial Attacks and Detection on Deep Learning-Based Plant Disease Identification

   https://doi.org/10.3390/app11041878  

   Luo, Zhirui; Li, Qingqing; Zheng, Jun (February 2021, Applied Sciences)

   null (Ed.)

   Transfer learning using pre-trained deep neural networks (DNNs) has been widely used for plant disease identification recently. However, pre-trained DNNs are susceptible to adversarial attacks which generate adversarial samples causing DNN models to make wrong predictions. Successful adversarial attacks on deep learning (DL)-based plant disease identification systems could result in a significant delay of treatments and huge economic losses. This paper is the first attempt to study adversarial attacks and detection on DL-based plant disease identification. Our results show that adversarial attacks with a small number of perturbations can dramatically degrade the performance of DNN models for plant disease identification. We also find that adversarial attacks can be effectively defended by using adversarial sample detection with an appropriate choice of features. Our work will serve as a basis for developing more robust DNN models for plant disease identification and guiding the defense against adversarial attacks. 

   more » « less
2. Learning Provably Robust Estimators for Inverse Problems via Jittering

   Krainovic, Anselm; Soltanolkotabi, Mahdi; Heckel, Reinhard (December 2023, Proceedings of Neural Information Processing Systems)

   Deep neural networks provide excellent performance for inverse problems such as denoising. However, neural networks can be sensitive to adversarial or worst-case perturbations. This raises the question of whether such networks can be trained efficiently to be worst-case robust. In this paper, we investigate whether jittering, a simple regularization technique that adds isotropic Gaussian noise during training, is effective for learning worst-case robust estimators for inverse problems. While well studied for prediction in classification tasks, the effectiveness of jittering for inverse problems has not been systematically investigated. In this paper, we present a novel analytical characterization of the optimal -worst-case robust estimator for linear denoising and show that jittering yields optimal robust denoisers. Furthermore, we examine jittering empirically via training deep neural networks (U-nets) for natural image denoising, deconvolution, and accelerated magnetic resonance imaging (MRI). The results show that jittering significantly enhances the worst-case robustness, but can be suboptimal for inverse problems beyond denoising. Moreover, our results imply that training on real data which often contains slight noise is somewhat robustness enhancing. 

   more » « less
3. A Robust Context-Based Deep Learning Approach for Highly Imbalanced Hyperspectral Classification

   https://doi.org/10.1155/2021/9923491  

   Ramirez Rochac, Juan F.; Zhang, Nian; Thompson, Lara A.; Deksissa, Tolessa (July 2021, Computational Intelligence and Neuroscience)

   Doulamis, Anastasios D. (Ed.)

   Hyperspectral imaging is an area of active research with many applications in remote sensing, mineral exploration, and environmental monitoring. Deep learning and, in particular, convolution-based approaches are the current state-of-the-art classification models. However, in the presence of noisy hyperspectral datasets, these deep convolutional neural networks underperform. In this paper, we proposed a feature augmentation approach to increase noise resistance in imbalanced hyperspectral classification. Our method calculates context-based features, and it uses a deep convolutional neuronet (DCN). We tested our proposed approach on the Pavia datasets and compared three models, DCN, PCA + DCN, and our context-based DCN, using the original datasets and the datasets plus noise. Our experimental results show that DCN and PCA + DCN perform well on the original datasets but not on the noisy datasets. Our robust context-based DCN was able to outperform others in the presence of noise and was able to maintain a comparable classification accuracy on clean hyperspectral images. 

   more » « less
4. Deep-Learning Based Estimation of Dielectrophoretic Force

   https://doi.org/10.3390/mi13010041  

   Ajala, Sunday; Jalajamony, Harikrishnan; Fernandez, Renny (January 2022, Micromachines)

   The ability to accurately quantify dielectrophoretic (DEP) force is critical in the development of high-efficiency microfluidic systems. This is the first reported work that combines a textile electrode-based DEP sensing system with deep learning in order to estimate the DEP forces invoked on microparticles. We demonstrate how our deep learning model can process micrographs of pearl chains of polystyrene (PS) microbeads to estimate the DEP forces experienced. Numerous images obtained from our experiments at varying input voltages were preprocessed and used to train three deep convolutional neural networks, namely AlexNet, MobileNetV2, and VGG19. The performances of all the models was tested for their validation accuracies. Models were also tested with adversarial images to evaluate performance in terms of classification accuracy and resilience as a result of noise, image blur, and contrast changes. The results indicated that our method is robust under unfavorable real-world settings, demonstrating that it can be used for the direct estimation of dielectrophoretic force in point-of-care settings. 

   more » « less
5. Comparing Detection Schemes for Adversarial Images against Deep Learning Models for Cancer Imaging

   https://doi.org/10.3390/cancers15051548  

   Joel, Marina Z.; Avesta, Arman; Yang, Daniel X.; Zhou, Jian-Ge; Omuro, Antonio; Herbst, Roy S.; Krumholz, Harlan M.; Aneja, Sanjay (March 2023, Cancers)

   Deep learning (DL) models have demonstrated state-of-the-art performance in the classification of diagnostic imaging in oncology. However, DL models for medical images can be compromised by adversarial images, where pixel values of input images are manipulated to deceive the DL model. To address this limitation, our study investigates the detectability of adversarial images in oncology using multiple detection schemes. Experiments were conducted on thoracic computed tomography (CT) scans, mammography, and brain magnetic resonance imaging (MRI). For each dataset we trained a convolutional neural network to classify the presence or absence of malignancy. We trained five DL and machine learning (ML)-based detection models and tested their performance in detecting adversarial images. Adversarial images generated using projected gradient descent (PGD) with a perturbation size of 0.004 were detected by the ResNet detection model with an accuracy of 100% for CT, 100% for mammogram, and 90.0% for MRI. Overall, adversarial images were detected with high accuracy in settings where adversarial perturbation was above set thresholds. Adversarial detection should be considered alongside adversarial training as a defense technique to protect DL models for cancer imaging classification from the threat of adversarial images. 

   more » « less