YOLSO: You Only Look Small Object

Small object detection is challenging and far from satisfactory. Most general object detectors suffer from two critical issues with small objects: (1) Feature extractor based on classification network cannot express the characteristics of small objects reasonably due to insufficient appearance information of targets and a large amount of background interference around them. (2) The detector requires a much higher location accuracy for small objects than for general objects. This paper proposes an effective and efficient small object detector YOLSO to address the above problems. For feature representation, we analyze the drawbacks in previous backbones and present a Half-Space Shortcut(HSSC) module to build a background-aware backbone. Furthermore, a coarse-to-fine Feature Pyramid Enhancement(FPE) module is introduced for layer-wise aggregation at a granular level to enhance the semantic discriminability. For loss function, we propose an exponential L1 loss to promote the convergence of regression, and a focal IOU loss to focus on prime samples with high classification confidence and high IOU. Both of them significantly improves the location accuracy of small objects. The proposed YOLSO sets state-of-the-art results on two typical small object datasets, MOCOD and VeDAI, at a speed of over 200 FPS. In the meantime, it also outperforms the baseline YOLOv3 by a wide margin on the common COCO dataset.     

Feature selection methods for root-cause analysis among top-level product attributes

Manufacturing companies not only strive to deliver flawless products but also monitor product failures in the field to identify potential quality issues. When product failures occur, quality engineers must identify the root cause to improve any affected product and process. This root-cause analysis can be supported by feature selection methods that identify relevant product attributes, such as manufacturing dates with an increased number of product failures. In this paper, we present different methods for feature selection and evaluate their ability to identify relevant product attributes in a root-cause analysis. First, we compile a list of feature selection methods. Then, we summarize the properties of product attributes in warranty case data and discuss these properties regarding the challenges they pose for machine learning algorithms. Next, we simulate datasets of warranty cases, which emulate these product properties. Finally, we compare the feature selection methods based on these simulated datasets. In the end, the univariate filter information gain is determined to be a suitable method for a wide range of applications. The comparison based on simulated data provides a more general result than other publications, which only focus on a single use case. Due to the generic nature of the simulated datasets, the results can be applied to various root-cause analysis processes in different quality management applications and provide a guideline for readers who wish to explore machine learning methods for their analysis of quality data.     

Breeding schemes with optimum-contribution selection or truncation selection for beef cattle destined for use on dairy females

We tested the hypothesis that the size of a beef cattle population destined for use on dairy females is smaller under optimum-contribution selection (OCS) than under truncation selection (TRS) at the same genetic gain (ΔG) and the same rate of inbreeding (ΔF). We used stochastic simulation to estimate true ΔG realized at a 0.005 ΔF in breeding schemes with OCS or TRS. The schemes for the beef cattle population also differed in the number of purebred offspring per dam and the total number of purebred offspring per generation. Dams of the next generation were exclusively selected among the one-year-old heifers. All dams were donors for embryo transfer and produced a maximum of 5 or 10 offspring. The total number of purebred offspring per generation was: 400, 800, 1,600 or 4,000 calves, and it was used as a measure of population size. Rate of inbreeding was predicted and controlled using pedigree relationships. Each OCS (TRS) scheme was simulated for 10 discrete generations and replicated 100 (200) times. The OCS scheme and the TRS scheme with a maximum of 10 offspring per dam required approximately 783 and 1,257 purebred offspring per generation to realize a true ΔG of €14 and a ΔF of 0.005 per generation. Schemes with a maximum of 5 offspring per dam required more purebred offspring per generation to realize a similar true ΔG and a similar ΔF. Our results show that OCS and multiple ovulation and embryo transfer act on selection intensity through different mechanisms to achieve fewer selection candidates and fewer selected sires and dams than under TRS at the same ΔG and a fixed ΔF. Therefore, we advocate the use of a breeding scheme with OCS and multiple ovulation and embryo transfer for beef cattle destined for use on dairy females because it is favorable both from an economic perspective and a carbon footprint perspective.     

Prediction of mode I fracture toughness of rock using linear multiple regression and gene expression programming

Prediction of mode I fracture toughness (K_IC) of rock is of significant importance in rock engineering analyses. In this study, linear multiple regression (LMR) and gene expression programming (GEP) methods were used to provide a reliable relationship to determine mode I fracture toughness of rock. The presented model was developed based on 60 datasets taken from the previous literature. To predict fracture parameters, three mechanical parameters of rock mass including uniaxial compressive strength (UCS), Brazilian tensile strength (BTS), and elastic modulus (E) have been selected as the input parameters. A cluster of data was collected and divided into two random groups of training and testing datasets. Then, different statistical linear and artificial intelligence based nonlinear analyses were conducted on the training data to provide a reliable prediction model of K_IC. These two predictive methods were then evaluated based on the testing data. To evaluate the efficiency of the proposed models for predicting the mode I fracture toughness of rock, various statistical indices including coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE) were utilized herein. In the case of testing datasets, the values of R², RMSE, and MAE for the GEP model were 0.87, 0.188, and 0.156, respectively, while they were 0.74, 0.473, and 0.223, respectively, for the LMR model. The results indicated that the selected GEP model delivered superior performance with a higher R² value and lower errors.     

Understanding the asymmetric perceptions of smartphone security from security feature perspective: A comparative study

Smartphones are being used and relied on by people more than ever before. The open connectivity brings with it great convenience and leads to a variety of risks that cannot be overlooked. Smartphone vendors, security policy designers, and security application providers have put a variety of practical efforts to secure smartphones, and researchers have conducted extensive research on threat sources, security techniques, and user security behaviors. Regrettably, smartphone users do not pay enough attention to mobile security, making many efforts futile. This study identifies this gap between technology affordance and user requirements, and attempts to investigate the asymmetric perceptions toward security features between developers and users, between users and users, as well as between different security features. These asymmetric perceptions include perceptions of quality, perceptions of importance, and perceptions of satisfaction. After scoping the range of smartphone security features, this study conducts an improved Kano-based method and exhaustively analyzes the 245 collected samples using correspondence analysis and importance satisfaction analysis. The 14 security features of the smartphone are divided into four Kano quality types and the perceived quality differences between developers and users are compared. Correspondence analysis is utilized to capture the relationship between the perceived importance of security features across different groups of respondents, and results of importance-satisfaction analysis provide the basis for the developmental path and resource reallocation strategy of security features. This article offers new insights for researchers as well as practitioners of smartphone security.     

Improved error performance for generalised spatial modulation with enhanced spectral efficiency

The existing analytical average bit error rate (ABER) expression of conventional generalised spatial modulation (CGSM) does not agree well with the Monte Carlo simulation results in the low signal‐to‐noise ratio (SNR) region. Hence, the first contribution of this paper is to derive a new and easy way to evaluate analytical ABER expression that improves the validation of the simulation results at low SNRs. Secondly, a novel system termed CGSM with enhanced spectral efficiency (CGSM‐ESE) is presented. This system is realised by applying a rotation angle to one of the two active transmit antennas. As a result, the overall spectral efficiency is increased by 1 bit/s/Hz when compared with the equivalent CGSM system. In order to validate the simulation results of CGSM‐ESE, the third contribution is to derive an analytical ABER expression. Finally, to improve the ABER performance of CGSM‐ESE, three link adaptation algorithms are developed. By assuming full knowledge of the channel at the receiver, the proposed algorithms select a subset of channel gain vector (CGV) pairs based on the Euclidean distance between all CGV pairs, CGV splitting, CGV amplitudes, or a combination of these.     

一种新的多渐消因子容积卡尔曼滤波

将强跟踪思想引入容积卡尔曼滤波（cubature Kalman filter，CKF），建立强跟踪CKF能有效克服CKF在模型不确定、状态突变等情况下，滤波性能下降的问题。通过分析现有多渐消因子计算方法，发现它们均只利用了协方差矩阵的对角线元素，并没有考虑各个状态之间的相关性，不能充分发挥多渐消因子的优势。为此，本文提出渐消因子矩阵，基于正交原理推导渐消因子矩阵的求解方法，提出多渐消因子强跟踪CKF算法。多渐消因子强跟踪CKF算法突破了传统多渐消因子为向量的限制，也不再要求渐消因子取值要大于1。仿真验证了算法具有更好的滤波精度何鲁棒性，能更好的满足工程应用的要求。     

基于梯度提升决策树的车轮轮缘厚度磨耗预测

轮对在列车走行过程中起着导向、承受以及传递载荷的作用，其踏面及轮缘磨耗对地铁列车运行安全性和钢轨的寿命都将产生重要影响。根据地铁列车车轮磨耗机理，分析车轮尺寸数据特点，针对轮缘厚度这一型面参数，基于梯度提升决策树算法构建轮缘厚度磨耗预测模型。在该模型的基础上，任意选取某轮对数据进行验证分析，结果表明：基于梯度提升决策树的轮对磨耗预测模型具有较好的预测精度，可预测出1~6个月的轮缘厚度变化趋势范围，预测时间范围较长，可为地铁维保部门对轮对的维修方式由状态修转为预防修提供指导性建议。     

基于深度学习的秀丽隐杆线虫显微图像分割方法

利用计算机实现自动、准确的秀丽隐杆线虫(C.elegans)的各项形态学参数分析,至关重要的是从显微图像上分割出线虫体态,但由于显微镜下的图像噪声较多,线虫边缘像素与周围环境相似,而且线虫的体态具有鞭毛和其他附着物需要分离,多方面因素导致设计一个鲁棒性的C.elegans分割算法仍然面临着挑战。针对这些问题,提出了一种基于深度学习的线虫分割方法,通过训练掩模区域卷积神经网络(Mask R-CNN)学习线虫形态特征实现自动分割。首先,通过改进多级特征池化将高级语义特征与低级边缘特征融合,结合大幅度软最大损失(LMSL)损失算法改进损失计算;然后,改进非极大值抑制;最后,引入全连接融合分支等方法对分割结果进行进一步优化。实验结果表明,相比原始的Mask R-CNN,该方法平均精确率(AP)提升了4.3个百分点,平均交并比(mIOU)提升了4个百分点。表明所提出的深度学习分割方法能够有效提高分割准确率,在显微图像中更加精确地分割出线虫体。     

基于词法特征的恶意域名快速检测算法

针对互联网中恶意域名攻击事件频发，现有域名检测方法实时性不强的问题，提出一种基于词法特征的恶意域名快速检测算法。该算法根据恶意域名的特点，首先将所有待测域名按照长度进行正则化处理后赋予权值；然后利用聚类算法将待测域名划分成多个小组，并利用改进的堆排序算法按照组内权值总和计算各域名小组优先级，根据优先级降序依次计算各域名小组中每一域名与黑名单上域名之间的编辑距离；最后依据编辑距离值快速判定恶意域名。算法运行结果表明，基于词法特征的恶意域名快速检测算法与单一使用域名语义和单一使用域名词法的恶意域名检测算法相比，准确率分别提高1.7%与2.5%，检测速率分别提高13.9%与6.8%，具有更高的准确率和实时性。