Deep neural network model is the core technology for sensors of the autonomous driving platform to perceive the external environment. Recent research have shown that it has a certain vulnerability. The artificial designed adversarial examples can make the DNN model output the wrong results. These adversarial examples not only exist in the digital world, but also in the physical world. At present, research on autonomous driving platform mainly focus on attacking a single sensor. In this paper, we introduce Camdar-adv, a method for generating image adversarial examples on three-dimensional (3D) objects, which could potentially lunch a multisensor attack toward the autonomous driving platforms. Specifically, with objects that can attack LiDAR sensors, a geometric transformation can be used to project their shape onto the twodimensional plane. Adversarial perturbations against optical image sensor could be added to the surface of the adversarial 3D objects precisely without changing its geometry. Test results on the open-source autonomous driving data set KITTI show that Camdar-adv can generate adversarial samples for the state of the art object detection model. From a fixed viewpoint, our method can achieve an attack success rate over 99%.

Background: To effectively make use of deep learning technology automatic feature extraction ability, and enhance the ability of depth learning method to learn and recognize features, this paper proposed a deep learning algorithm combining Deep Convolutional Neural Network (DCNN) trained with an improved cost function and Support Vector Machine (SVM). Methods: The class separation information, which explicitly facilitates intra-class compactness and interclass separability in the process of learning features, is added to an improved cost function as a regularization term to enhance the feature extraction ability of DCNN. Then the improved DCNN is applied to learn the features of SAR images. Finally, SVM is utilized to map the features into output labels. Results: Experiments are performed on SAR image data in Moving and Stationary Target Acquisition and Recognition (MSTAR) database. The experiment results prove the effectiveness of our method, achieving an average accuracy of 99% on ten types of targets, some variants, and some articulated targets. Conclusion: It proves that our method is effective and CNN enjoys a certain potential to be applied in SAR image target recognition. Keywords synthetic aperture radar (SAR) images • automatic target recognition (ATR) • deep convolutional neural Network (DCNN) • support vector machine (SVM) • class separation information

No abstract

Abstract:The deep convolution neural network (CNN), which has prominent advantages in feature learning, can learn and extract features from data automatically. Existing polarimetric synthetic aperture radar (PolSAR) image classification methods based on the CNN only consider the polarization information of the image, instead of incorporating the image's spatial information. In this paper, a novel method based on a dual-branch deep convolution neural network (Dual-CNN) is proposed to realize the classification of PolSAR images. The proposed method is built on two deep CNNs: one is used to extract the polarization features from the 6-channel real matrix (6Ch) which is derived from the complex coherency matrix. The other is utilized to extract the spatial features of a Pauli RGB (Red Green Blue) image. These extracted features are first combined into a fully connected layer sharing the polarization and spatial property. Then, the Softmax classifier is employed to classify these features. The experiments are conducted on the Airborne Synthetic Aperture Radar (AIRSAR) data of Flevoland and the results show that the classification accuracy on 14 types of land cover is up to 98.56%. Such results are promising in comparison with other state-of-the-art methods.

The high mutation rates of infectious bronchitis virus (IBV) pose economic threats to the poultry industry. In order to track the genetic evolutionary of IBV isolates circulating in yellow chickens, we continued to conduct the genetic analyses of the structural genes S1, E, M, and N from 64 IBV isolates in southern China during 2009–2017. The results showed that the dominant genotypes based on the four genes had changed when compared with those during 1985–2008. Based on the S1 gene phylogenetic tree, LX4-type (GI-19) was the most dominant genotype, which was different from that during 1985–2008. The second most dominant genotype was LDT3-A-type, but this genotype disappeared after 2012. New-type 1 (GVI-1) isolates showed increasing tendency and there were four aa (QKEP) located in the hypervariable region (HVR) III and one aa (S) insertion in all the New-type 1 isolates. Both the analyses of amino acid entropy and molecular evolutionary rate revealed that the variations from large to small were S1, E, M, and N. Purifying selection was detected in the S1, E, M, and N gene proteins, which was different from the positive selection during 1985–2008. Six isolates were confirmed to be recombinants, possibly generated from a vaccine virus of the 4/91-type or LDT3-A-type and a circulating virus. The estimated times for the most recent common ancestors based on the S1, E, M, and N genes were the years of 1744, 1893, 1940, and 1945, respectively. Bayesian skyline analysis revealed a sharp decrease in genetic diversity of all the four structural genes after 2010 and since late 2015, the viral population rapidly rose. In conclusion, the IBVs circulating in southern China over the past decade have experienced a remarkable change in genetic diversity, dominant genotypes, and selection pressure, indicating the importance of permanent monitoring of circulating strains and the urgency for developing new vaccines to counteract the emerging LX4-type and New-type IBVs.

No abstract

No abstract

No abstract