Countermeasure against blinding attack for single-photon detectors in quantum key distribution

Quantum key distribution(QKD),rooted in quantum mechanics,offers information-theoretic security.However,practi-cal systems open security threats due to imperfections,notably bright-light blinding attacks targeting single-photon detectors.Here,we propose a concise,robust defense strategy for protecting single-photon detectors in QKD systems against blinding attacks.Our strategy uses a dual approach:detecting the bias current of the avalanche photodiode(APD)to defend against con-tinuous-wave blinding attacks,and monitoring the avalanche amplitude to protect against pulsed blinding attacks.By integrat-ing these two branches,the proposed solution effectively identifies and mitigates a wide range of bright light injection attempts,significantly enhancing the resilience of QKD systems against various bright-light blinding attacks.This method forti-fies the safeguards of quantum communications and offers a crucial contribution to the field of quantum information security.

Superconducting Nanowire Single Photon Detector with Optical Cavity

Increasing the detection efficiency （DE） is a hot issue in the development of the superconducting nanowire single photon detector （SNSPD）. In this work, a cavity-integrated structure coupled to the SNSPD is used to enhance the light absorption of nanowire. Ultra-thin Nb films are successfully prepared by magnetron sputtering, which are used to fabricate Nb/Al SNSPD with the curve of lOOnm and the square area of 4 × 4μm2 by sputtering and the lift-off method. To characterize the optical and electrical performance of the cavity-integrated SNSPD, a reliable cryogenic research system is built up based on a He3 system. To satisfy the need of light coupling, a packaging structure with collimator is conducted. Both DE and the dark count rates increase with lb. It is also found that the DE of SNSPD with cavities can be up to 0.17% at the temperature of 0.7K under the infrared light of 1550nm, which is obviously higher than that of the SNSPD directly fabricated upon silicon without any cavity structure.

Detection efficiency characteristics of free-running InGaAs/InP single photon detector using passive quenching active reset IC

InGaAs/InP avalanche photodiodes （APD） are rarely used in a free-running regime for near-infrared single photon detection. In order to overcome the detrimental afterpulsing, we demonstrate a passive quenching active reset integrated circuit. Taking advantage of the inherent fast passive quenching process and active reset to reduce reset time, the integrated circuit is useful for reducing afterpulses and is also area-efficient. We investigate the free-running single photon detector＇s afterpulsing effect, de-trapping time, dark count rate, and photon detection efficiency, and also compare with gated regime operation. After correction for deadtime and afterpulse, we find that the passive quenching active reset free-running single photon detector＇s performance is consistent with gated operation.

The analysis of the integral gated mode single photon detector

This paper critically analyses and simulates the circuit configuration of the integral gated mode single photon detector which is proposed for eliminating the transient spikes problem of conventional gated mode single photon detector. The relationship between the values of the circuit elements and the effect of transient spikes cancellation has been obtained. With particular emphasis, the bias voltage of the avalanche photodiode and the output signal voltage of the integrator have been calculated. The obtained analysis results indicate that the output signal voltage of the integrator only relates to the total quantity of electricity of the avalanche charges by choosing the correct values of the circuit elements and integral time interval. These results can be used to optimize the performance of single photon detectors and provide guides for the design of single photon detectors.

The intensity detection of single-photon detectors based on photon counting probability density statistics

Single-photon detectors possess the ultra-high sensitivity, but they cannot directly respond to signal intensity. Conven- tional methods adopt sampling gates with fixed width and count the triggered number of sampling gates, which is capable of obtaining photon counting probability to estimate the echo signal intensity. In this paper, we not only count the number of triggered sampling gates, but also record the triggered time position of photon counting pulses. The photon counting probability density distribution is obtained through the statistics of a series of the triggered time positions. Then Minimum Variance Unbiased Estimation （MVUE） method is used to estimate the echo signal intensity. Compared with conventional methods, this method can improve the estimation accuracy of echo signal intensity due to the acquisition of more detected information. Finally, a proof-of-principle laboratory system is established. The estimation accuracy of echo signal intensity is discussed and a high accuracy intensity image is acquired under low-light level environments.

Performance of a superconducting single photon detector with nano-antenna

The performance of single-photon detectors can be enhanced by using nano-antenna. The characteristics of the superconducting nano-wire single-photon detector with cavity plus anti-reflect coating and specially designed nano~ antenna is analysed. The photon collection efficiency of the detector is enhanced without damaging the detector＇s speed, thus getting rid of the dilemma of speed and efficiency. The characteristics of nano-antenna are discussed, such as the position and the effect of the active area, and the best result is given. The photon collection efficiency is increased by 92 times compared with that of existing detectors.

Implementation of a single quad MS detector in routine QC analysis of peptide drugs

A newly developed single quad mass spectrometry （MS） detector was coupled to a ultra-high perfor- mance liquid chromatography （UPLC） system and implemented in the routine quality control （QC） and impurity analysis of four therapeutic peptides, namely bleomycin sulfate, tyrothricin, vancomycin HCl and bacitracin, which were selected given their multi-component drug nature and their closely struc- turally related impurity profiles. The QC and impurity profiling results obtained using the ultra-high performance liquid chromatography ultraviolet/mass spectrometry （UPLC-UV/MS） detection system were analyzed against the results obtained using traditional high performance liquid chromatography- ultraviolet detection （HPLC-UV） methods derived from pharmacopoeial methods. In general, the used stationary phases of sub-2 μm particle （UPLC） technology resulted in lower limits of detection and higher resolution separations, which resulted in more detected impurities and shorter overall run times con- trasting the traditional HPLC columns. Moreover, online coupling with a single quad MS detector allowed direct peak identification of the main compounds as well as small impurities, hereby increasing the information content without the need of reference standards.

Statistical analysis of the temporal single-photon response of superconducting nanowire single photon detection

A new method to study the transient detection efficiency （DE） and pulse amplitude of superconducting nanowire single photon detectors （SNSPD） during the current recovery process is proposed -- statistically analyzing the single photon response under photon illumination with a high repetition rate. The transient DE results match well with the DEs deduced from the static current dependence of DE combined with the waveform of a single-photon detection event. This proves that static measurement results can be used to analyze the transient current recovery process after a detection event. The results are relevant for understanding the current recovery process of SNSPDs after a detection event and for determining the counting rate of SNSPDs.

Detailed calibration of the PI-LCX：1300 high performance single photon counting hard x-ray CCD camera

X-ray charge-coupled-device（CCD） camera working in single photon counting mode is a type of x-ray spectrometer with high-sensitivity and superior signal-to-noise performance. In this study, two single photon counting CCD cameras with the same mode（model： PI-LCX： 1300） are calibrated with quasi-monochromatic x-rays from radioactive sources and a conventional x-ray tube. The details of the CCD response to x-rays are analyzed by using a computer program of multi-pixel analyzing and event-distinguishing capability. The detection efficiency, energy resolution, fraction of multi-pixel events each as a function of x-ray energy, and consistence of two CCD cameras are obtained. The calibrated detection efficiency is consistent with the detection efficiency from Monte Carlo calculations with XOP program. When the multi-pixel event analysis is applied, the CCDs may be used to measure x-rays up to 60 ke V with good energy resolution（E /？E ≈ 100 at60 ke V）. The difference in detection efficiency between two CCD cameras is small（5.6% at 5.89 ke V）, but the difference in fraction of the single pixel event between them is much larger（25% at 8.04 ke V）. The obtained small relative error of detection efficiency（2.4% at 5.89 ke V） makes the high accurate measurement of x-ray yield possible in the laser plasma interaction studies. Based on the discrete calibration results, the calculated detection efficiency with XOP can be used for the whole range of 5 ke V–30 ke V.

一种基于机器视觉的精准注意力追踪系统

针对学生注意力分配困难和对学习影响等问题,提出一种基于机器视觉的精准注意力追踪系统。该系统包括图像采集装置和精准的注意力追踪算法。图像采集装置可以获得更清晰的眼部区域图像。瞳孔中心定位算法用轻量级的MobileNet v3替换VGG16(visual geometry group network),采用两级特征融合和中心关键点预测技术,提高了检测速度和准确率。该算法检测速度可达36帧/s,准确率为97.42%。视线追踪算法旨在解决头部偏移的影响,实现对视线的精确追踪。研发了一款面向学龄儿童的阅读认知评价交互软件。该软件利用采集到的视线坐标计算相关眼动指标,再通过心理学理论分析建模来评估学龄儿童的思维认知能力,为心理学和教育学相关领域研究提供了参考和借鉴。

Deep Learning Based Face Detection and Identification of Criminal Suspects

Occurrence of crimes has been on the constant rise despite the emerging discoveries and advancements in the technological field in the past decade.One of the most tedious tasks is to track a suspect once a crime is committed.As most of the crimes are committed by individuals who have a history of felonies,it is essential for a monitoring system that does not just detect the person’s face who has committed the crime,but also their identity.Hence,a smart criminal detection and identification system that makes use of the OpenCV Deep Neural Network(DNN)model which employs a Single Shot Multibox Detector for detection of face and an auto-encoder model in which the encoder part is used for matching the captured facial images with the criminals has been proposed.After detection and extraction of the face in the image by face cropping,the captured face is then compared with the images in the CriminalDatabase.The comparison is performed by calculating the similarity value between each pair of images that are obtained by using the Cosine Similarity metric.After plotting the values in a graph to find the threshold value,we conclude that the confidence rate of the encoder model is 0.75 and above.

基于特征融合的SSD视觉小目标检测

针对SSD算法在检测目标过程中对小目标检测效果差的缺陷,提出了特征融合的SSD方法。该方法充分融合深浅层特征信息以提升网络模型对小目标的检测能力,为更好地检测小目标,将先验框尺寸相对原图比列进行调整,同时对SSD模型相应超参数值进行调整。实验结果表明,检测精度mAP较SSD提高3.4个百分点,对小目标Bottle、Chair、Plant检测精度分别提升8.7个百分点、3.4个百分点和7.1个百分点。检测精度mAP较当前一系列性能优异的目标检测算法有显著提高。通过拓展实验进一步证明改进算法成功检测到了大多数SSD算法没有检测到的小目标,提高了平均检测准确率。

基于密集模块与特征融合的SSD目标检测算法

通过对原SSD(Single Shot Multibox Detector)模型的研究与分析,针对其对小目标检测能力较弱的问题,提出了一种基于密集模块与特征融合操作的改进模型。该模型以Inception-ResNet-V2与DenseNet为基础,吸取了inception模块中稀疏连接与密集网络中密集连接的研究思路,将两种方法融合在一起,提出了Inception-Dense特征提取结构。在多尺度检测的部分,借鉴并改进了特征金字塔的特征融合模块来加强对中小目标的检测能力。根据改进模型及实验数据集的相关特性,对默认框的映射机制也进行了重新设定。结果表明:该方法在Kitti数据集上的平均测试精确度(mAP)为83.8%;识别率相比于原SSD模型的72.8%,提升了11个百分点。FPS方面也有接近38%的提升,从原来的39提升到了54。

基于ECA-SSD模型的汽车零件缺陷检测

汽车零件对汽车外观、性能以及安全性都有重大影响。由于汽车零件数量大、体积小、对精度要求高,因此对零件检测的精度和速度都有一定的要求。本文利用图像处理技术,以SSD模型为基础,将其中的VGG模块用深度可分离卷积和线性瓶颈倒残差结构替换,并引入避免降维的局部跨通道交互有效的注意力机制ECA模块,在减少模型参数运算量的同时,适当增加通道以提高模型精度,并将注意力放在图像目标上,忽略背景带来的干扰,实现快速又准确的汽车零件缺陷检测。利用本文模型对上汽提供的汽车零件外壁缺陷进行检测,实验结果表明,模型大小仅为15.9 MB,mAP为94.64%,检测每张图片时间为0.013 s,满足汽车工业上的速度和精度的需求。对比性研究表明,本文模型检测精度和速度以及大小较其他目标检测算法VGG-SSD、MobileNetv2-SSD、YOLO v3等有一定的提高和改善。

Quest towards ultimate performance in superconducting nanowire single photon detectors

In 2007, superconducting nanowire single photon detectors （SSPD or SNSPD） [ 1 ] made an outstanding impact in the field of quantum information technology by demonstrating quan- tum key distribution （QKD） over a 200-kin optical fiber with a 42-dB optical loss using a practical SNSPD system [2]. This successful demonstration was realized thanks to its extremely low dark count rate （DCR） of a few Hz and short timing jitter of 60 ps, while the system detection efficiency （SDE） showed a poor value of 0.7% at a wavelength of 1550 nm.

A method for robust TV logo detection

A robust TV logo detection method based on the modified single shot multibox detector (SSD) is presented. Unlike most other existing methods which can only detect the TV logo from video frames, the proposed method can also detect the TV logo from photo pictures taken by smartphones or other smart terminals. Firstly, using a simple and effective way of collecting and labelling TV logo, a large-scale TV logo dataset used to train the detection model is built. Then, parameters and loss function of SSD are modified to make it more suitable for the task of TV logo detection. Moreover, a soft-NMS algorithm is introduced to remove the redundant overlapping boxes and obtain the final output box. And also an approach for hard example mining is designed to improve the detection accuracy. Finally, extensive comparison experiments are carried out which take into consideration different image resolutions, logo positions and environmental factors existing in real-world applications. Experimental results demonstrate that the proposed method achieve superior performances in robustness compared to other state-of-the-art methods.

Polarization independent superconducting nanowire detector with high-detection efficiency

The superconducting nanowire single photon detector（SNSPD） draws much attention because of its attractive performance at ultra violet, visible, and nearinfrared wavelengths, and it can be widespread in quantum information technologies. However, how to increase the absorption which can dramatically increase the quantum efficiency of the SNSPD is still a top research issue. In this study, the effect of incident medium and cavity material on the optical absorptance of cavity-integrated SNSPDs was systematically investigated using finite-element method. The simulation results demonstrate that for photons polarized parallel to nanowire orientation, even though the maximum absorptance of the nanowire is insensitive to cavity material,it does increase when the refractive index of incident medium decreases. For perpendicularly polarized photons, both incident medium and cavity material play significant roles,and the absorptance curves get closer to the parallel case as the refractive index of cavity material increases. Based on these results, two cavity-integrated SNSPDs with frontillumination structure which can enhance the absorptance for both parallel and perpendicular photons are proposed.Finally, a design to realize polarization-independent SNSPDs with high absorptance is presented.

Automated classification and detection of multiple pavement distress images based on deep learning

To achieve automatic,fast,efficient and high-precision pavement distress classification and detection,road surface distress image classification and detection models based on deep learning are trained.First,a pavement distress image dataset is built,including 9017pictures with distress,and 9620 pictures without distress.These pictures were captured from 4 asphalt highways of 3 provinces in China.In each pavement distress image,there exists one or more types of distress,including alligator crack,longitudinal crack,block crack,transverse crack,pothole and patch.The distresses are labeled by a rectangle bounding box on the pictures.Then ResNet networks and VGG networks are used respectively as binary classification models for distressed and non-distressed imagines classification,and as multi-label classification models for six types of distress classification.Training techniques,such as data augmentation,batch normalization,dropout,momentum,weight decay,transfer learning,and discriminative learning rate are used in training the model.Among the 4 CNNs considered in this study,namely ResNet 34 and 50,and VGG 16 and 19,for the binary classification,ResNet 50 has the highest Accuracy of 96.243%,Precision of 95.183%,and ResNet 34 has the highest Recall of 97.824%,and F2 score of 97.052%.For multi-label classification,ResNet 50 has the best performance,with the highest Accuracy of 90.257%,higher than 90%required by the Chinese standard(JTG H20-2018)for road distresses detection,F2 score-82.231%,and Precision-76.509%,and ResNet34 has the highest Recall of 87.32%.To locate and quantify the distress areas in the images,the single shot multibox detector(SSD)model is developed,in which the ResNet 50 is used as the base network to extract features.When the intersection over union(IoU)is set to 0,0.25,0.50,0.75,the mean average precision(mAP)of the model are found to be 74.881%,50.511%,28.432%,3.969%,respectively.

基于空间-通道注意力的改进SSD目标检测算法

目标检测的任务是精确识别,有效定位出图像中目标物体,且预定义其类别。针对主流目标检测(single shot multibox detector,SSD)算法存在小目标检测准确度不高,检测效率较低等问题,提出一种基于空间-通道注意力机制的SSD目标检测算法(spatial and channel single shot multibox detector,SC_SSD)。通过在SSD深层网络引入空间-通道注意力机制增强高层特征图语义信息,提高算法获取目标物体的细节与位置信息的能力,从而降低漏检率及误检率,并提高小目标物体检测的准确度。此外,利用MobileNetV2中的深度可分离卷积对SSD主干网络(visual geometry group network,VGG-16)进行剪枝处理,降低参数量,从而减少训练与检测的时间。在PASCAL VOC2007数据集上进行实验,本文算法检测的精确度与速度分别为78.9%与59.4 Fps,比SSD算法提升了3.2%与26.7 Fps,满足实时性需求。算法也优于相比较的其他算法,是一种有效可行的目标检测算法。

Multi-block SSD based on small object detection for UAV railway scene surveillance

A method of multi-block Single Shot Multi Box Detector(SSD)based on small object detection is proposed to the railway scene of unmanned aerial vehicle surveillance.To address the limitation of small object detection,a multi-block SSD mechanism,which consists of three steps,is designed.First,the original input images are segmented into several overlapped patches.Second,each patch is separately fed into an SSD to detect the objects.Third,the patches are merged together through two stages.In the first stage,the truncated object of the sub-layer detection result is spliced.In the second stage,a sub-layer suppression and filtering algorithm applying the concept of non-maximum suppression is utilized to remove the overlapped boxes of sub-layers.The boxes that are not detected in the main-layer are retained.In addition,no sufficient labeled training samples of railway circumstance are available,thereby hindering the deployment of SSD.A two-stage training strategy leveraging to transfer learning is adopted to solve this issue.The deep learning model is preliminarily trained using labeled data of numerous auxiliaries,and then it is refined using only a few samples of railway scene.A railway spot in China,which is easily damaged by landslides,is investigated as a case study.Experimental results show that the proposed multi-block SSD method produces an overall accuracy of 96.6%and obtains an improvement of up to 9.2%compared with the traditional SSD.