GAN-based data augmentation of prohibited item X-ray images in security inspection

Convolutional neural networks (CNNs) based methods for automatic discriminant of prohibited items in X-ray images attract attention increasingly. However, it is difficult to train a reliable CNN model using the available X-ray security image databases, since they are not enough in sample quantity and diversity. Recently, generative adversarial network (GAN) has been widely used in image generation and regarded as a power model for data augmentation. In this paper, we propose a data augmentation method for X-ray prohibited item images based on GAN. First, the network structure and loss function of the self-attention generative adversarial network (SAGAN) are improved to generate the realistic X-ray prohibited item images. Then, the images generated by our model are evaluated using GAN-train and GAN-test. Experimental results of GAN-train and GAN-test are 99.91% and 98.82% respectively. It implies that our model can enlarge the X-ray prohibited item image database effectively.     

Automatic detection of prohibited items with small size in X-ray images

In this paper, we focus on the detection of prohibited items with small size, and establish an automatic detection model based on feature fusion single shot multibox detector (FSSD) architecture. Two modifications are carried out to improve the detection accuracy. Firstly, the semantic enrichment module (SEM) with dilated convolution is applied to extract the low level feature with strong semantic information. Secondly, a residual module (Res) with residual blocks is added in the multibox detection architecture in order to extract more adequate features for target detection. The simulation results have demonstrated a better performance of the proposed detection model for prohibited items with small size compared with the state-of-the-arts.     

Compact diffraction grating laser wavemeter for cold atom experiments

We present an innovative and practical scheme of building a miniaturized wavemeter, with the advantages of low cost, high reliability and simple structure. Through a calibration test by a 780 nm external cavity diode laser (ECDL), the results show that our system gets a wavelength resolution of better than 1 pm, measurement accuracy of better than 2 pm (corresponding to a frequency of 1 GHz), and a measurement range of 8.5 nm. Finally, the multi-mode comparison test between our system and a commercial spectrum analyzer further indicates the high-precision, miniaturization and low cost of the proposed system, which shows that it is particularly suitable for ECDL and atom cooling and trapping experiments. The system design, experimental results and conclusions are of definite significance as a fine reference for other ranges of wavelength.     

A robust color image watermarking algorithm against rotation attacks

A robust digital watermarking algorithm is proposed based on quaternion wavelet transform (QWT) and discrete cosine transform (DCT) for copyright protection of color images. The luminance component Y of a host color image in YIQ space is decomposed by QWT, and then the coefficients of four low-frequency subbands are transformed by DCT. An original binary watermark scrambled by Arnold map and iterated sine chaotic system is embedded into the mid-frequency DCT coefficients of the subbands. In order to improve the performance of the proposed algorithm against rotation attacks, a rotation detection scheme is implemented before watermark extracting. The experimental results demonstrate that the proposed watermarking scheme shows strong robustness not only against common image processing attacks but also against arbitrary rotation attacks.     

An accurate pose measurement method of workpiece based on rapid extraction of local feature points

Ceramic sanitary products with complex curved surfaces are generally fragile and difficult to clamp. If the industrial robot is utilized to realize the automatic grinding of such products, the precise positioning of the product is required firstly. In this paper, an accurate pose measurement system for complex curved surface parts is designed by point cloud registration algorithm. In order to improve the stability of the system, this paper combines the advantages of normal vector features and fast point feature histogram (FPFH) features, and proposes a point cloud registration algorithm based on the rapid extraction of local feature points. Experimental results verify that the improved algorithm has improved both efficiency and accuracy, and the system can effectively achieve accurate positioning of products.     

Multi-scale detector optimized for small target

The effectiveness of deep learning networks in detecting small objects is limited, thereby posing challenges in addressing practical object detection tasks. In this research, we propose a small object detection model that operates at multiple scales. The model incorporates a multi-level bidirectional pyramid structure, which integrates deep and shallow networks to simultaneously preserve intricate local details and augment global features. Moreover, a dedicated multi-scale detection head is inte...     

Image restoration of finger-vein networks based on encoder-decoder model

Finger-vein recognition is widely applied on access control system due to the high user acceptance and convince. Improving the integrity of finger-vein is helpful for increasing the finger-vein recognition accuracy. During the process of finger-vein imaging, foreign objects may be attached on fingers, which directly affects the integrity of finger-vein images. In order to effectively extract finger-vein networks, the integrity of venous networks is still not ideal after preprocessing of finger vein images. In this paper, we propose a novel deep learning based image restoration method to improve the integrity of finger-vein networks. First, a region detecting method based on adaptive threshold is presented to locate the incomplete region. Next, an encoder-decoder model is used to restore the venous networks of the finger-vein images. Then we analyze the restoration results using several different methods. Experimental results show that the proposed method is effective to restore the venous networks of the finger-vein images.     

Design of zero cross correlation variable weight codes for multimedia services based on magic square in SAC-OCDMA systems

A variable weight address code based on spectrum amplitude coding (SAC) is proposed for optical code division multiple access (OCDMA) networks to support different quality of service (QoS) requirements of different services. The zero cross-correlation magic square variable weight optical orthogonal code (ZMS-VWOOC) proposed in this paper has great flexibility in terms of code weight and number of users. Zero cross-correlation can eliminate the influence of multiple access interference (MAI) and reduce the system complexity. Numerical results show that ZMS-VWOOC can provide better quality of service than similar codes. Numerical results for a ZMS-VWOOC OCDMA network designed for triple-play services operating at 0.622 Gbit/s, 2 Gbit/s, 2.5 Gbit/s and 3 Gbit/s are considered.     

Real-time human blood pressure measurement based on laser self-mixing interferometry with extreme learning machine

In this paper, we present a method based on self-mixing interferometry combing extreme learning machine for real-time human blood pressure measurement. A signal processing method based on wavelet transform is applied to extract reversion point in the self-mixing interference signal, thus the pulse wave profile is successfully reconstructed. Considering the blood pressure values are intrinsically related to characteristic parameters of the pulse wave, 80 samples from the MIMIC-II database are used to train the extreme learning machine blood pressure model. In the experiment, 15 measured samples of pulse wave signal are used as the prediction sets. The results show that the errors of systolic and diastolic blood pressure are both within 5 mmHg compared with that by the Coriolis method.     

Q-switched giant pulsed erbium-doped all-fiber laser withV2ZnC MAX phase saturable absorber

MXenes, drawn from MAX phases, are special two-dimensional substances with numerous advantages in nonlinear optics, specifically in giant and ultrashort pulsed-laser applications. Ti3C2Tx and Ti2CTx nanosheets however rapidly deteriorate under ambient conditions, limiting their applications. This paper demonstrates how excellent modulation depth of one of the MAX phase compounds vanadium zinc carbide (V2ZnC) makes it a brilliant saturable absorber (SA) in passively Q-switched all-fiber pulsed lasers, integrated such that a 16.73-µm V2ZnC-polyvinyl alcohol (PVA) thin film acts as SA in the laser. Saturable and non-saturable absorptions were found to be 13.2% and 10.47%, while saturation optical intensity and modulation depth were 6.25 kW/cm2 and 12.43%, respectively, illustrating the optical nonlinearity. The superiority of MAX-PVA, fabricated in four distinct ratios, was demonstrated by the fact that it self-starts a giant pulsed laser at pump power as low as 22.5 mW and firmly accomplished 120.6 kHz repetition rate with a pulse width of 2.08 µs. It is a fine SA for the use of pulsed-laser production using all-fiber laser due to fabrication simplicity and great optical, thermophysical, and mechanical qualities.     

Hybrid plasmonic label-free multi-analyte refractive index sensor

We propose a label-free refractive index sensor based on hybrid plasmonic resonator which consists of silver split-ring resonator and photonic waveguide. The finite difference time domain evaluation of the design exhibit strong field confinement at the center of the ring and introduces tunable and sensitive notches in the transmission spectrum. The planar tunable architecture which performs well over the range of micro fluid detection, holds the promise of developing multi-analyte label-free biosensors and compactness towards a complete on-chip integrated sensing system. The performance of the proposed refractive index sensor is evaluated by placing different analytes such as saline water and ethanol at the center of the hybrid plasmonic ring which exhibits sensitivity of 847.50 nm.RIU-1 with a figure of merit of 563.25 RIU-1.     

A prohibited items identification approach based on semantic segmentation

Deep learning (DL) based semantic segmentation methods can extract object information including category, location and shape. In this paper, the identification of prohibited items is regarded as a task of semantic segmentation, and proposes a universal model with automatic identification of prohibited items. This model has two improvements based on the general semantic segmentation network. Firstly, the N-type encoding structure is applied to enlarge the receptive field of the network aiming at reducing the misclassification. Secondly, consider the lack of surface texture in X-ray security images. Inspired by feature reuse in Densenet, shallow semantic information is reused to improve the segmentation accuracy. With the use of this model, when using input images of size 512×512, we could achieve 0.783 mean intersection over union (mIoU) for a seven-class object recognition problem.     

X光安检图像多尺度违禁品检测

安检是保障人民生命财产安全的第一道防线,智能安检是安检行业未来发展的必然趋势。X光安检图像存在背景复杂、违禁品尺度多样以及相互遮挡现象,导致传统的目标检测算法无法获得满意的效果。本文在一阶段目标检测网络SSD框架的基础上,提出了适用于X光安检图像多尺度违禁品检测网络——非对称卷积多视野神经网络ACMNet(Asymmetrical Convolution Multi-View Neural Network)。检测网络增加了三个模块:小卷积非对称模块(Asymmetrical Tiny Convolution Module,ATM)、空洞多视野卷积模块(Dilated Convolution Multi-View Module,DCM)、多尺度特征图融合策略(Fusion strategy of multi-scale feature map,MF)。 ATM学习到的细节特征有助于小尺度违禁品的识别;DCM通过提供局部与全局之间的上下文特征信息来解决遮挡问题;MF则是通过融合高、低层特征图以提高模型在背景干扰情况下违禁品的检测精度。在仿真实验中,采用X光安检领域公开的数据集与自建的数据集,ACMNet在精确度上取得了令人满意的效果。      

基于语义分割的多目标违禁品识别算法

基于深度学习的语义分割算法可以实现安检违禁品自动识别,并获得违禁品的位置、类别及形状信息。但传统的语义分割算法在面对违禁品尺寸不一且目标多样的识别任务时表现较差。针对该问题,本文提出了一种基于语义分割技术的多目标违禁品识别算法。编码阶段,设计使用空洞空间金字塔卷积模块（Atrous Spatial Pyramid Convolution Block, ASPC）,提升网络对于特征图多尺度信息的挖掘能力。同时引入注意力机制,对ASPC模块的特征提取过程进行监督,进一步提升模块的特征提取能力。解码阶段,受U-Net模型启发,采用逐级上采样操作,同时加入1×1卷积实现通道降维,减少计算量,提升模型运行速度。实验结果显示,本文提出的算法在多目标违禁品识别任务中表现良好,平均交并比（mIoU）得分78.62,处理单张图片用时（Time）68ms。      

基于图匹配网络的小样本违禁物品分割算法

自动化安检技术是维护公共安全、提升安检效率的一项有效措施。在实际场景中很难获得充足的违禁品标注样本用于神经网络的训练,并且在不同场景和安全级别下违禁品的类别也有所不同。为解决基于神经网络的违禁品检测方法所面临的样本不均衡问题,以及避免模型在分割新的违禁品类别时需重新训练的现象,文中提出一种基于图匹配网络的小样本违禁物品分割算法。文中模型将测试图像与参考图像并行输入到图匹配网络中,并根据匹配结果从测试图像中分割出违禁品。所设计的图匹配模块不仅从图间节点的相似性考虑匹配问题,并利用DeepEMD算法建立全局概念,进一步提高测试图和参考图的匹配结果。在SIXray数据集和Xray-PI数据集上的实验表明:本模型在单样本分割任务中得到36.4%和51.2%的类平均交并比,分别比目前先进的单样本分割方法提高2.5%和2.3%。由此表明所设计的算法能有效提升小样本X光图像分割算法的精确度。     

Optimal Security Proofs for Full Domain Hash,Revisited

RSA Full Domain Hash (RSA-FDH) is a digital signature scheme, secure against chosen message attacks in the random oracle model. The best known security reduction from the RSA assumption is non-tight, i.e., it loses a factor of \(q_s\), where \(q_s\) is the number of signature queries made by the adversary. It was furthermore proven by Coron (Advances in cryptology—EUROCRYPT 2002, Lecture notes in computer science, vol 2332. Springer, Berlin, pp 272–287, 2002) that a security loss of \(q_s\) is optimal and cannot possibly be improved. In this work, we uncover a subtle flaw in Coron’s impossibility result. Concretely, we show that it only holds if the underlying trapdoor permutation is certified. Since it is well known that the RSA trapdoor permutation is (for all practical parameters) not certified, this renders Coron’s impossibility result moot for RSA-FDH. Motivated by this, we revisit the question whether there is a tight security proof for RSA-FDH. Concretely, we give a new tight security reduction from a stronger assumption, the Phi-Hiding assumption introduced by Cachin et al. (Advances in Cryptology—EUROCRYPT’99. Lecture notes in computer science, vol 1592. Springer, Berlin, pp 402–414, 1999). This justifies the choice of smaller parameters in RSA-FDH, as it is commonly used in practice. All of our results (positive and negative) extend to the probabilistic signature scheme PSS (with message recovery).     

On the Impact of Quality of Protection in Wireless Local Area Networks with IP Mobility

Wireless local area networks (LANs) are vulnerable to malicious attacks due to their shared medium in unlicensed frequency spectrum, thus requiring security features for a variety of applications even at the cost of quality of service (QoS). However, there is very little work on investigating to what extent system performance is affected by security configurations with respect to mobility scenarios, heterogeneous networks, and different applications. In order to exploit the full potential of existing security solutions, we present a detailed experimental study to demonstrate the impacts of security features on performance by integrating cross-layer security protocols in a wireless LAN testbed with IP mobility. We introduce a quality of protection (QoP) model to indicate the benefits of security protocols and then measure the performance cost of security protocols in terms of authentication time, cryptographic overhead and throughput. Our measurements demonstrate that the effects of security protocols on QoS parameters span a wide range; for example, authentication time is between 0.11 and 6.28 s, which can potentially affect packet loss dramatically. We also find that for the same security protocol throughput in non-roaming scenarios can be up to two times higher than that in roaming scenarios. However, some protocols are robust against mobility with little variation in system performance; thus, it is possible to provision steady service by choosing security protocols when users’ mobility pattern is unknown. Furthermore, we provide observations on cross-layer security protocols and suggestions to the design of future security protocols for real-time services in wireless LANs.

Efficient Cryptosystems From $$\mathbf{2}^{{\varvec{k}}}$$-th Power Residue Symbols

Goldwasser and Micali (J Comput Syst Sci 28(2):270–299, 1984) highlighted the importance of randomizing the plaintext for public-key encryption and introduced the notion of semantic security. They also realized a cryptosystem meeting this security notion under the standard complexity assumption of deciding quadratic residuosity modulo a composite number. The Goldwasser–Micali cryptosystem is simple and elegant but is quite wasteful in bandwidth when encrypting large messages. A number of works followed to address this issue and proposed various modifications. This paper revisits the original Goldwasser–Micali cryptosystem using \(2^k\)-th power residue symbols. The so-obtained cryptosystems appear as a very natural generalization for \(k \ge 2\) (the case \(k=1\) corresponds exactly to the Goldwasser–Micali cryptosystem). Advantageously, they are efficient in both bandwidth and speed; in particular, they allow for fast decryption. Further, the cryptosystems described in this paper inherit the useful features of the original cryptosystem (like its homomorphic property) and are shown to be secure under a similar complexity assumption. As a prominent application, this paper describes an efficient lossy trapdoor function-based thereon.     

Restoration in wavelength-routed optical networks by means of ant colony optimization

Because of the distributed control of the network, the dynamic nature of the traffic and the unpredictability of a failure event, the flexibility and robustness of ant colony optimization (ACO) make it a suitable candidate for provisioning lightpaths in an optical network. In this work, we propose a fault-tolerant dynamic routing and wavelength assignment (RWA) algorithm based on the ACO framework, presenting its integration into the Generalized multi-protocol label switching (GMPLS) control plane. By simulating two different scenarios, we demonstrate the effectiveness of this algorithm when a single link or node failure occurs.