An Empirical Study of Adversarial Examples on Remote Sensing Image Scene Classification

Remote sensing image (RSI) scene classification is the foundation and important technology of ground object detection, land use management and geographic analysis. During recent years, convolutional neural networks (CNNs) have achieved significant success and are widely applied in RSI scene classification. However, crafted images that serve as adversarial examples can potentially fool CNNs with high confidence and are hard for human eyes to interpret. For the increasing security and robust requirements of RSI scene classification, the adversarial example problem poses a serious problem for the classification results derived from systems using CNN models, which has not been fully recognized by previous research. In this study, to explore the properties of adversarial examples of RSI scene classification, we create different scenarios by testing two major attack algorithms (i.e., the fast gradient sign method (FGSM) and basic iterative method (BIM)) trained on different RSI benchmark datasets to fool CNNs (i.e., InceptionV1, ResNet and a simple CNN). In the experiment, our results show that CNNs of RSI scene classification are also vulnerable to adversarial examples, and some of them have a fooling rate of over 80%. These adversarial examples are affected by the architecture of CNNs and the type of RSI dataset. InceptionV1 has a fooling rate of less than 5%, which is lower than the others. Adversarial examples generated on the UCM dataset are easier than other datasets. Importantly, we also find that the classes of adversarial examples have an attack selectivity property. Misclassifications of adversarial examples of RSIs are related to the similarity of the original classes in the CNN feature space. Attack selectivity reveals potential classes of adversarial examples and provides insights into the design of defensive algorithms in future research.

Deep neural networks have made great achievements in remote sensing image analyses; however, previous studies have shown that deep neural networks exhibit incredible vulnerability to adversarial examples, which raises concerns about regional safety and production safety. In this paper, we propose an adversarial denoising method based on latent representation guidance for remote sensing image scene classification. In the training phase, we train a variational autoencoder to reconstruct the data using only the clean dataset. At test time, we first calculate the normalized mutual information between the reconstructed image using the variational autoencoder and the reference image as denoised by a discrete cosine transform. The reconstructed image is selectively utilized according to the result of the image quality assessment. Then, the latent representation of the current image is iteratively updated according to the reconstruction loss so as to gradually eliminate the influence of adversarial noise. Because the training of the denoiser only involves clean data, the proposed method is more robust against unknown adversarial noise. Experimental results on the scene classification dataset show the effectiveness of the proposed method. Furthermore, the method achieves better robust accuracy compared with state-of-the-art adversarial defense methods in image classification tasks.

Synthetic aperture radar (SAR) image classification is a challenging problem due to the complex imaging mechanism as well as the random speckle noise, which affects radar image interpretation. Recently, deep neural networks (DNNs) have been shown to outperform previous state-of-the-art techniques in computer vision tasks owing to their ability to learn relevant features from the data. However, the fragility of these models has received far less academic attention in the remote sensing community, which limits our understanding of the security of remote sensing image classification models. To explore the basic characteristic of adversarial examples of SAR images, we compare several mainstream adversarial methods and evaluate the securities of used DNNs from the perspective of attention. We subsequently perform other attempts. The experimental results provide data support and an effective reference for the defense capabilities of various DNNs regarding attack in SAR image classification models.

Some recent articles have revealed that synthetic aperture radar automatic target recognition (SAR-ATR) models based on deep learning are vulnerable to the attacks of adversarial examples and cause security problems. The adversarial attack can make a deep convolutional neural network (CNN)-based SAR-ATR system output the intended wrong label predictions by adding small adversarial perturbations to the SAR images. The existing optimization-based adversarial attack methods generate adversarial examples by minimizing the mean-squared reconstruction error, causing smooth target edge and blurry weak scattering centers in SAR images. In this paper, we build a UNet-generative adversarial network (GAN) to refine the generation of the SAR-ATR models’ adversarial examples. The UNet learns the separable features of the targets and generates the adversarial examples of SAR images. The GAN makes the generated adversarial examples approximate to real SAR images (with sharp target edge and explicit weak scattering centers) and improves the generation efficiency. We carry out abundant experiments using the proposed adversarial attack algorithm to fool the SAR-ATR models based on several advanced CNNs, which are trained on the measured SAR images of the ground vehicle targets. The quantitative and qualitative results demonstrate the high-quality adversarial example generation and excellent attack effectiveness and efficiency improvement.

The methods for remote sensing image (RSI) scene classification based on deep convolutional neural networks (DCNNs) have achieved prominent success. However, confronted with adversarial examples obtained by adding imperceptible perturbations to clean images, the great vulnerability of DCNNs makes it worth exploring effective defense methods. To date, numerous countermeasures for adversarial examples have been proposed, but how to improve the defensive ability for unknown attacks still to be answered. To address this issue, in this article, we propose an effective defense framework specified for RSI scene classification, named perturbation-seeking generative adversarial networks (PSGANs). In brief, a new training framework is designed to train the classifier by introducing the examples generated during the image reconstruction process, in addition to clean examples and adversarial ones. These generated examples can be random kinds of unknown attacks during training and thus are utilized to eliminate the blind spots of a classifier. To assist the proposed training framework, a reconstruction method is developed. First, instead of modeling the distribution of clean examples, we model the distributions of the perturbations added in adversarial examples. Second, to make a tradeoff between the diversity of the reconstructed examples and the optimization of PSGAN, a scale factor named seeking radius is introduced to scale the generated perturbations before they are subtracted by the given adversarial examples. Comprehensive and extensive experimental results on three widely used benchmarks for RSI scene classification demonstrate the great effectiveness of PSGAN when faced with both known and unknown attacks. Our source code is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/xuxiangsun/PSGAN</uri> .

This paper presents a novel approach based on the direct use of deep neural networks to approximate wavelet sub-bands for remote sensing (RS) image scene classification in the JPEG 2000 compressed domain. The proposed approach consists of two main steps. The first step aims to approximate the finer level wavelet sub-bands. To this end, we introduce a novel Deep Neural Network approach that utilizes the coarser level binary decoded wavelet sub-bands to approximate the finer level wavelet sub-bands (the image itself) through a series of deconvolutional layers. The second step aims to describe the high-level semantic content of the approximated wavelet sub- bands and to perform scene classification based on the learnt descriptors. This is achieved by: i) a series of convolutional layers for the extraction of descriptors which models the approximated sub-bands; and ii) fully connected layers for the RS image scene classification. Then, we introduce a loss function that allows to learn the parameters of both steps in an end-to-end trainable and unified neural network. The proposed approach requires only the coarser level wavelet sub-bands as input and thus minimizes the amount of decompression applied to the compressed RS images. Experimental results show the effectiveness of the proposed approach in terms of classification accuracy and reduced computational time when compared to the conventional use of Convolutional Neural Networks within the JPEG 2000 compressed domain.

Deep neural network-based synthetic aperture radar (SAR) automatic target recognition (ATR) systems are susceptible to attack by adversarial examples, which leads to misclassification by the SAR ATR system, resulting in theoretical model robustness problems and security problems in practice. Inspired by optical images, current SAR ATR adversarial example generation is performed in the image domain. However, the imaging principle of SAR images is based on the imaging of the echo signals interacting between the SAR and objects. Generating adversarial examples only in the image domain cannot change the physical world to achieve adversarial attacks. To solve these problems, this article proposes a framework for generating SAR adversarial examples in a 3D physical scene. First, adversarial attacks are implemented in the 2D image space, and the perturbation in the image space is converted into simulated rays that constitute SAR images through backpropagation optimization methods. The mapping between the simulated rays constituting SAR images and the 3D model is established through coordinate transformation, and point correspondence to triangular faces and intensity values to texture parameters are established. Thus, the simulated rays constituting SAR images are mapped to the 3D model, and the perturbation in the 2D image space is converted back to the 3D physical space to obtain the position and intensity of the perturbation in the 3D physical space, thereby achieving physical adversarial attacks. The experimental results show that our attack method can effectively perform SAR adversarial attacks in the physical world. In the digital world, we achieved an average fooling rate of up to 99.02% for three objects in six classification networks. In the physical world, we achieved an average fooling rate of up to 97.87% for these objects, with a certain degree of transferability across the six different network architectures. To the best of our knowledge, this is the first work to implement physical attacks in a full physical simulation condition. Our research establishes a theoretical foundation for the future concealment of SAR targets in practical settings and offers valuable insights for enhancing the attack and defense capabilities of subsequent DNNs in SAR ATR systems.

Despite deep neural networks (DNNs) have been widely applied in remote sensing (RS) scene classification and achieved satisfying performance, the vulnerability of DNNs towards adversarial examples significantly degrades their performance. Moreover, the relatively limited labeled samples of RS scene classification make DNNs more likely to overfit, leading to weak generalizability and noise sensitivity. This may result in DNNs being more vulnerable to adversarial examples. Consequently, the defense of adversarial examples is of crucial importance to improve both the generalizability and robustness of DNNs in the RS scene classification task. However, few studies have been conducted on defense for RS scene classification, especially ignoring the intrinsic characteristics of RS images. In this paper, an effective defense framework for RS scene classification, named reconstruction-assisted and distance-optimized adversarial training (RDAT), is proposed to defend adversarial examples. In order to solve the problems caused by high interclass similarity, a distance-optimized (DO) strategy is designed for adversarial training to strengthen the learning of underfitting content, increase the interclass distance, and improve the robustness of the networks. Furthermore, in order to generate high quality samples for adversarial training, a reconstruction-assisted (RA) block is proposed to eliminate adversarial perturbations in adversarial examples. Specifically, in this block, by swin transformer (SwinT) block and multi-scale convolution (MSC) block, SwinT-MSC-UNet (SMUNet) is constructed to fully extract global and multi-scale local features to adapt to the characteristics of RS images with large variance of ground object scales. Extensive experiments on the benchmark datasets, i.e., UC Merced (UCM) and Aerial Image Dataset (AID), have demonstrate that the proposed RDAT can effectively resist multiple adversarial attacks and yield superior results than other defense methods for RS scene classification.

Remote sensing image scene classification has a high application value in the agricultural, military, as well as other fields. A large amount of remote sensing data is obtained every day. After learning the new batch data, scene classification algorithms based on deep learning face the problem of catastrophic forgetting, that is, they cannot maintain the performance of the old batch data. Therefore, it has become more and more important to ensure that the scene classification model has the ability of continual learning, that is, to learn new batch data without forgetting the performance of the old batch data. However, the existing remote sensing image scene classification datasets all use static benchmarks and lack the standard to divide the datasets into a number of sequential learning training batches, which largely limits the development of continual learning in remote sensing image scene classification. First, this study gives the criteria for training batches that have been partitioned into three continual learning scenarios, and proposes a large-scale remote sensing image scene classification database called the Continual Learning Benchmark for Remote Sensing (CLRS). The goal of CLRS is to help develop state-of-the-art continual learning algorithms in the field of remote sensing image scene classification. In addition, in this paper, a new method of constructing a large-scale remote sensing image classification database based on the target detection pretrained model is proposed, which can effectively reduce manual annotations. Finally, several mainstream continual learning methods are tested and analyzed under three continual learning scenarios, and the results can be used as a baseline for future work.

Deep Neural Networks (DNN) have strong capabilities of memories, feature identifications and automatic analyses, solving various complex problems. However, DNN classifiers have obvious fragility that adding several unnoticeable perturbations to the original examples will lead to the errors in the classifier identification. In the field of communications, the adversarial examples will greatly reduce the accuracy of the signal identification, causing great information security risks. Considering the adversarial examples pose a serious threat to the security of the DNN models, studying their generation mechanisms and testing their attack effects are critical to ensuring the information security of the communication networks. This paper will study the generation of the adversarial examples and the influences of the adversarial examples on the accuracy of the DNN-based communication signal identification. Meanwhile, this paper will study the influences of the adversarial examples under the white-box models and black-box models, and explore the adversarial attack influences of the factors such as perturbation levels and iterative steps. The insights of this study would be helpful for ensuring the security of information networks and designing robust DNN communication networks.

Recent research has revealed that the deep neural network (DNN)-based synthetic-aperture radar (SAR) automatic target recognition (ATR) techniques are vulnerable to adversarial examples, which poses significant security risks for their deployment in real-world systems. At the same time, the adversarial examples often exhibit transferability across DNN models, whereby when they are generated on the surrogate model they can also attack other target models. As the significant property in black-box scenarios, transferability has been enhanced by various methods, among which input transformations have demonstrated excellent effectiveness. However, we find that existing transformations suffer from limited enhancement of transferability due to the unique imaging mechanism and scattering characteristics of SAR images. To overcome this issue, we propose a novel method called intra-class transformations and inter-class nonlinear fusion attack (ITINFA). It enhances transferability from two perspectives: intra-class single image transformations and inter-class multiple images fusion. The intra-class transformations module utilizes a series of diverse transformations that align with the intrinsic characteristics of SAR images to obtain a more stable gradient update direction and prevent the adversarial examples from overfitting the surrogate model. The inter-class fusion strategy incorporates the information from other categories in a nonlinear manner, effectively enhances the feature fusion effect, and guides the misclassification of adversarial examples. Extensive experiments on the MSTAR dataset and SEN1-2 dataset demonstrate that ITINFA exhibits significantly better transferability than the existing transfer-based methods, with the average transfer attack success rate increases exceeding 8% for single models and over 4% for ensemble models.

The classification of aerial scenes has been extensively studied as the basic work of remote sensing image processing and interpretation. However, the performance of remote sensing image scene classification based on deep neural networks is limited by the number of labeled samples. In order to alleviate the demand for massive labeled samples, various methods have been proposed to apply semi-supervised learning to train the classifier using labeled and unlabeled samples. However, considering the complex contextual relationship and huge spatial differences, the existing semi-supervised learning methods bring different degrees of incorrectly labeled samples when pseudo-labeling unlabeled data. In particular, when the number of labeled samples is small, it affects the generalization performance of the model. In this article, we propose a novel semi-supervised learning method with early labeled and small loss selection. First, the model learns the characteristics of simple samples in the early stage and uses multiple early models to screen out a small number of unlabeled samples for pseudo-labeling based on this characteristic. Then, the model is trained in a semi-supervised manner by combining labeled samples, pseudo-labeled samples, and unlabeled samples. In the training process of the model, small loss selection is used to further eliminate some of the noisy labeled samples to improve the recognition accuracy of the model. Finally, in order to verify the effectiveness of the proposed method, it is compared with several state-of-the-art semi-supervised classification methods. The results show that when there are only a few labeled samples in remote sensing image scene classification, our method is always better than previous methods.

Deep learning architectures have emerged as powerful function approximators in a broad spectrum of complex representation learning tasks, such as, computer vision, natural language processing and collaborative filtering. These architectures bear a high potential to learn the intrinsic structure of data and extract valuable insights. Despite the surge in the development of state-of-the-art intelligent systems using the deep neural networks (DNNs), these systems have found to be vulnerable to adversarial examples produced by adding a small-magnitude of perturbations. Such adversarial examples are adept at misleading the DNN classifiers. In the past, different attack strategies have been proposed to produce adversarial examples in the digital, physical, and transform domain, but the likelihood to generate perceptually realistic adversarial examples require more research efforts. In this paper, we present a novel approach to produce adversarial examples by combining the single-shot fast gradient sign method (FGSM) and spatial, as well as, transform domain image processing techniques. The resulted perturbations neutralize the impact of low-intensity based regions, thus, instilling the noise only in the selective high-intensity regions of the input image. While combining the customized perturbation with one-step FGSM perturbation in an un-targeted black-box attack scenario, the proposed approach successfully fools state-of-the-art DNN classifiers with 99% adversarial examples being misclassified on the ImageNet validation dataset.

Many companies offer automatic speech recognition or Speech-to-Text APIs for use in diverse applications. However, audio classification algorithms trained with deep neural networks (DNNs) can sometimes misclassify adversarial examples, posing a significant threat to critical applications. In this paper, we present a novel way to create adversarial audio examples using a genetic algorithm. Our algorithm creates adversarial examples by iteratively adding perturbations to the original audio signal. Unlike most white-box adversarial example generations, our approach does not require knowledge about the target DNN's model and parameters (black-box) and heavy computational power of GPU resources. To show the feasibility of the proposed idea, we implement a tool called, Foolgle, using a genetic algorithm that performs untargeted attacks to create adversarial audio examples and evaluate those with the state-of-the-art Google Cloud Speech-to-Text API. Our preliminary experiment results show that Foolgle deceives the API with a success probability of 86%.

With the rapid development of Deep Neural Networks (DNNs), remote sensing image analysis has achieved significant progress in scene classification, object detection, and semantic segmentation. However, the increasing deployment of DNNs in real-world remote sensing applications exposes them to adversarial risks. A critical challenge in this domain is the poor visual stealth of black-box adversarial examples, perturbations are often easily perceived by humans, limiting their practicality. This issue stems from the distribution mismatch between visually salient regions and model-sensitive regions in remote sensing images. The current approaches predominantly focus on improving attack success rates while neglecting such heterogeneity, resulting in redundant and conspicuous perturbations. To address this, we propose a Dual-Region Hierarchical Attack (DRHA) that improves stealth and attack performance by using the different roles of visually salient and model-sensitive regions. Shallow perturbations are applied to salient areas to preserve perceptual similarity, while directional perturbations guided by integrated gradients target high-contribution regions to enhance model deception. Our method leverages superpixel segmentation and attribution analysis to localize perturbation regions precisely. Experiments on the UCM and AID datasets show that DRHA improves average success rates by 10%–32% over baseline methods, while maintaining superior stealth and perturbation sparsity. These results demonstrate the effectiveness of region-aware attack design in constructing imperceptible and transferable adversarial examples for remote sensing tasks.