基于YOLOv3和EPnP算法的多药盒姿态估计

针对机械臂药盒抓取操作中对药盒定位和姿态估计的要求,提出一种基于YOLOv3深度学习算法和EPnP算法相结合的多药盒姿态估计方法,此方法主要分为多药盒定位和姿态估计两部分;首先通过YOLOv3算法实现药盒的快速精确定位,并通过定位框分割出单个药盒;然后进行特征提取和特征匹配并估计单应矩阵;通过单应矩阵的透视矩阵变换求得药盒平面4个角点的像素坐标并作为EPnP求解所需的2D点,结合药盒先验尺寸信息在相机坐标系下构建药盒对应的3D点坐标以实现药盒姿态求解;通过结合OptiTrack系统设计了药盒姿态精度对比实验,结果表明,该算法充分发挥了YOLOv3算法兼具快速性和准确性的优势,并且具有良好的姿态估计精度,总体算法速度达到15 FPS,药盒姿态估计平均误差小于0.5°。     

Adaptive video object proposals by a context-aware model

Most previous works focus on image object proposals while few on video object proposals. Besides, the existing explorations about video object proposals mainly concentrate on localizing the dominant object. In this paper, we aim at exploring a uniform framework for proposing multi-objects in videos no matter they are in the foreground or background. The method is derived from image object proposals, and makes best use of video characteristics. To achieve this task, we propose an adaptive context-aware model for video object proposals. First, spatial candidate windows are generated by the image method for acquiring the adequate bounding box samples. Temporal boxes are calculated by the motion based mapping. Considering the mapping loss, we define a box confidence coefficient contributing to keeping the proposal consistency and restraining the motion blur. The output proposal bounding boxes are ranked based on the scores calculated by the weighted scoring system. The proposed method is separately evaluated on the proposed multi-object dataset and the public dataset. The results compared with several state-of-the-arts show that our method has the most satisfactory overall performance for multi-object proposals in videos.     

Partial Word Order Freezing in Dutch

Dutch allows for variation as to whether the first position in the sentence is occupied by the subject or by some other constituent, such as the direct object. In particular situations, however, this commonly observed variation in word order is ‘frozen’ and only the subject appears in first position. We hypothesize that this partial freezing of word order in Dutch can be explained from the dependence of the speaker’s choice of word order on the hearer’s interpretation of this word order. A formal model of this interaction between the speaker’s perspective and the hearer’s perspective is presented in terms of bidirectional Optimality Theory. Empirical predictions of this model regarding the interaction between word order and definiteness are confirmed by a quantitative corpus study.     

Effects of box features on spine loading during warehouse order selecting.

Low back disorders in distribution centres or warehouses have been identified as an area of elevated risk in many industries. The task of an order selector requires workers manually to lift boxes from storage bins to a mobile pallet. This study explored the effect of box features and box location when lifting from a pallet in a storage bin upon spine loading. Ten experienced warehouse workers were asked to lift boxes from a pallet while the size, weight, handle features and location of the box on a pallet were changed. An EMG-assisted model was employed to assess spine compression, lateral shear and anterior-posterior shear during the lifts. The position from which the worker lifted a box on a pallet had the most profound effect on spine loading while the lower level of the pallet represented the greatest loadings on the spine. Box weight did not appear to be a feasible means of controlling spine loading unless its position on the pallet could also be controlled. The inclusion of handles had an effect similar to reducing the box weight by 4.5 kg, whereas box size did not effectively affect spine loading. The mechanisms by which these factors affect spine loading are discussed.     

一种改善双射S盒非线性度的方法

S盒是分组密码中重要的非线性部件,S盒的密码性质直接影响了密码算法的安全性,一个好的S盒要求有较高的非线性度。William Millan曾给出一个能改善双射S盒非线性度的Hill Climbing算法,它通过交换S盒的两个输出向量来提高S盒的非线性度直到非线性度达到一个局部最优值,本文在此基础上研究了如何同时改变S盒的四个输出向量的位置来更大程度的提高S盒的非线性度。     

Effects of carrying methods and box handles on two-person team carrying capacity for females

This study used a psychophysical approach to examine the effects of carrying methods and the presence or absence of box handles on the maximum acceptable weight carried and resulting responses (heart rate and rating of perceived exertion) in a two-person carrying task. After training, 16 female subjects performed a two-person carrying task at knuckle height for an 8-h work period. Each subject performed 4 different carrying combinations two times. The independent variables were carrying methods (parallel and tandem walking) and box handles (with and without handles). For comparison with two-person carrying, the subjects also performed one-person carrying. The results showed that the maximum acceptable weight carried (MAWC), heart rate (HR), and rating of perceived exertion (RPE) were significantly affected by the presence of box handles. However, the subjects' MAWC, HR, and RPE values were not significantly influenced by the carrying methods. The test-retest reliability of the psychophysical approach was 0.945. The carrying efficiency of two-person carrying was 96.2% of the one-person carrying method. In general, the use of box with handles allows the subjects to carry a higher MAWC (with lower HR and RPE) compared to carrying boxes without handles.     

Effect of initial horizontal object position on peak L5/S1 moments in manual lifting is dependent on task type and familiarity with alternative lifting strategies

This study investigated whether the effects of initial horizontal object position on peak L5/S1 total moment (PTM) are affected by task type or familiarity with alternative lifting strategies during manual lifting. Nine subjects lifted low-lying boxes from far and close initial horizontal positions in a typical laboratory lifting task (without any transportation of the load) and in a more realistic lifting task in which the box was transported to a location at a few metres distance. Subsequently, subjects were familiarised with alternative lifting strategies (e.g. shifting and tilting) and they then repeated the more realistic lifting task. Compared with the typical laboratory lifting task, the more realistic lifting task resulted in 6% larger PTMs for the close-positioned box. Familiarisation with alternative lifting techniques resulted in a 10% reduction in PTMs for the far-positioned box. As a result, the effect of initial horizontal box position on PTMs was smaller for the more realistic lifting task than for the typical laboratory lifting task and vanished after familiarisation with alternative lifting strategies. STATEMENT OF RELEVANCE: This study showed that the effect of horizontal box position on peak L5/S1 moments is dependent on the type of lifting task (comparing a typical laboratory simulated lifting task with a more realistic task involving carrying the load for a short distance) and familiarity with alternative lifting strategies. Therefore, it is recommended that back loading should be evaluated in a realistic simulation of the work situation or at the workplace itself.     

主动迁移学习的海上任意方向船只目标检测

在基于深度学习的遥感图像目标检测任务中,船只目标通常呈现出任意方向排列的特性,而常见的水平框目标检测算法一般不能满足此类场景的应用需求。因此本文在单阶段Anchor-Free目标检测器CenterNet的基础上加入旋转角度预测分支,使其能输出旋转边界框,以用于海上船只目标的检测。同时针对海上船只遥感数据集仅有水平边界框标注,无法直接适用于旋转框目标检测,且人工手动标注旋转框标签成本较高的问题,提出一种主动迁移学习的旋转框标签生成方法。首先,提出一种水平框-旋转框约束筛选算法,通过水平真值边界框来对旋转预测框进行监督约束,筛选出检测精度较高的图像加入训练集,然后通过迭代这一过程筛选出更多的图像,最后通过标签类别匹配,完成对数据集的旋转框自动化标注工作。本文最终对海上船只遥感图像数据集BDCI中约65.59%的图片进行旋转框标注,并手动标注部分未标注的图片作为测试集,将本文方法标注的图片作为训练集进行验证,评估指标AP50达到90.41%,高于其他旋转框检测器,从而表明本文方法的有效性。     

A group theoretic approach to construct cryptographically strong substitution boxes

In this paper, we present a method to construct a substitution box used in encryption applications. The proposed algorithm for the construction of substitution box relies on the linear fractional transform method. The design methodology is simple, while the confusion-creating ability of the new substitution box is complex. The strength of the proposed substitution box is evaluated, and an insight is provided to quantify the confusion-creating ability. In addition, tests are performed to assess the vulnerability of the encrypted data to algebraic and statistical attacks. The substitution box is critically analyzed by strict avalanche criterion, bit independent criterion, differential approximation probability test, linear approximation probability test, non-linearity test, and majority logic criterion. The performance of the proposed substitution box is also compared with those of some of the well-known counterparts including AES, APA, Gray, S₈, Skipjack, Xyi, and prime of residue substitution boxes. It is apparent that the performance, in terms of confusion-creating ability, of the new substitution box is better than those of some of the existing non-linear components used in encryption systems. The majority logic criterion is applied to these substitution boxes to further evaluate the strength and usefulness.     

Parallel Transformation of Bounding Volume Hierarchies into Oriented Bounding Box Trees

Oriented bounding box (OBB) hierarchies can be used instead of hierarchies based on axis-aligned bounding boxes (AABB), providing tighter fitting to the underlying geometric structures and resulting in improved interference tests, such as ray-geometry intersections. In this paper, we present a method for the fast, parallel transformation of an existing bounding volume hierarchy (BVH), based on AABBs, into a hierarchy based on oriented bounding boxes. To this end, we parallelise a high-quality OBB extraction algorithm from the literature to operate as a standalone OBB estimator and further extend it to efficiently build an OBB hierarchy in a bottom up manner. This agglomerative approach allows for fast parallel execution and the formation of arbitrary, high-quality OBBs in bounding volume hierarchies. The method is fully implemented on the GPU and extensively evaluated with ray intersections.     

Effect of initial horizontal object position on peak L5/S1 moments in manual lifting is dependent on task type and familiarity with alternative lifting strategies

This study investigated whether the effects of initial horizontal object position on peak L5/S1 total moment (PTM) are affected by task type or familiarity with alternative lifting strategies during manual lifting. Nine subjects lifted low-lying boxes from far and close initial horizontal positions in a typical laboratory lifting task (without any transportation of the load) and in a more realistic lifting task in which the box was transported to a location at a few metres distance. Subsequently, subjects were familiarised with alternative lifting strategies (e.g. shifting and tilting) and they then repeated the more realistic lifting task. Compared with the typical laboratory lifting task, the more realistic lifting task resulted in 6% larger PTMs for the close-positioned box. Familiarisation with alternative lifting techniques resulted in a 10% reduction in PTMs for the far-positioned box. As a result, the effect of initial horizontal box position on PTMs was smaller for the more realistic lifting task than for the typical laboratory lifting task and vanished after familiarisation with alternative lifting strategies.

Statement of Relevance:This study showed that the effect of horizontal box position on peak L5/S1 moments is dependent on the type of lifting task (comparing a typical laboratory simulated lifting task with a more realistic task involving carrying the load for a short distance) and familiarity with alternative lifting strategies. Therefore, it is recommended that back loading should be evaluated in a realistic simulation of the work situation or at the workplace itself.     

Kinetic analysis of manual lifting activities: Part II—Biomechanical analysis of task variables

This paper, the second of a series of two papers, presents the results of biomechanical analyses of task variables in manual lifting activities. The three-dimensional dynamic biomechanical model, presented in part I was used to analyze compressive and shear forces generated during symmetrical and asymmetrical lifting, lifting boxes with or without handles, and lifting loads in different size boxes (defined by the box dimension in the sagittal plane). The measured ground reaction forces were also analyzed for the effects of these task variables. The results indicated that even though low-weights are accepted for lifting when lifting loads asymmetrically or in bigger boxes or when handling boxes without handles, the spinal stresses generated are, in general, significantly higher than when lifting loads symmetrically or in compart boxes or when handling boxes with handles. At the maximum acceptable weights of lift, the compressive forces generated were observed to be at least 30% to 50% lower than the compressive failure limit of the spinal structure.     

A survey of the optimal handle position for boxes with different sizes and manual handling positions

Handles on objects are very important for enhancing the safety and efficiency of manual handling for people who use them. In this study, four different prototype boxes with auxiliary handles were designed to determine the optimal handle position of a box based on the evaluated user preferences and body part discomfort (BPD). Twenty male students participated in the experiment. Likert-5 point summated rating was applied to evaluate user preferences for the provided boxes with handles in upper, middle, and lower positions, in four different sizes and manual handling positions. Ten additional subjects were asked to indicate their BPD on a body chart after performing a similar experiment. The results show that the subjects preferred the upper part of the handle on a small box regardless of handling position; while the mid to upper parts of the handle on a big box were preferred for handling above the waist height. BPD also indicated that an upper handle was less stressful for a relatively smaller box than a big one; and mid to upper handles were less comfortable for a big box. The optimal handle positions depending on box size and handling position were suggested based on the results of the evaluation. It is thus recommended that a box provides a handle according to its relevant position, depending on size and manual handling condition, to reduce the musculoskeletal stress and in turn to increase user satisfaction.     

Improved non-maximum suppression for object detection using harmony search algorithm

Non-maximum suppression (NMS) plays a key role in many modern object detectors. It is responsible to remove detection boxes that cover the same object. NMS greedily selects the detection box with maximum score; other detection boxes are suppressed when the degree of overlap between these detection boxes and the selected box exceeds a predefined threshold. Such a strategy easily retain some false positives, and it limits the ability of NMS to perceive nearby objects in cluttered scenes. This paper proposes an effective method combining harmony search algorithm and NMS to alleviate this problem. This method regards the task of NMS as a combination optimization problem. It seeks final detection boxes under the guidance of an objective function. NMS is applied to each harmony to remove imprecise detection boxes, and the remaining boxes are used to calculate the fitness value. The remaining detection boxes in a harmony with highest fitness value are chosen as the final detection results. The standard Pattern Analysis, Statistical Modeling and Computational Learning Visual Object Classes dataset and the Microsoft Common Objects in Context dataset are used in all of the experiments. The proposed method is applied to two popular detection networks, namely Faster Region-based Convolutional Neural Networks and Region-based Fully Convolutional Networks. The experimental results show that the proposed method improves the average precision of these two detection networks. Moreover, the location performance and average recall of these two detectors are also improved.     

基于中心点搜索的无锚框全卷积孪生跟踪器

为解决孪生网络跟踪器鲁棒性差的问题,重新设计了孪生网络跟踪器的分类与回归分支,提出一种基于像素上直接预测方式的高鲁棒性跟踪算法—无锚框全卷积孪生跟踪器(Anchor-free fully convolutional siamese tracker,AFST).目前高性能的跟踪算法,如SiamRPN、SiamRPN++、...     

Symmetric and asymmetric manual materials handling. Part 1: Physiology and psychophysics

To study the physiological and psychophysical costs of symmetric and asymmetric manual materials handling, two tasks were performed by 30 industrial subjects. In both tasks, box weight and handle position were varied. The symmetric task, lifting and lowering between floor and conveyor, showed handles to be beneficial. The asymmetric task was palletizing and depalletizing 36 boxes between a pallet and a conveyor. Both palletizing and depalletizing proved strenuous for females with heart rates exceeding 140b/min. All handle positions were better than No Handles, but the best handle position changed from asymmetric for 9 kg boxes to symmetric for 13 kg boxes. The effect of handles was equivalent to a weight change of 1-2 kg for Heart Rate and Rated Perceived Exertion, but much higher (2-14 kg) for Body Part Discomfort measures.     

确定任意形状物体最小包围盒的一种方法

最小包围盒广泛应用于碰撞检测、模具分型设计、产品包装设计以及图像处理、模式识别等领域。根据不同的应用场合对包围盒的不同要求,提出了一种简单、实用的确定任意物体最小包围盒的方法,利用现有的CAD软件本身具有确定轴向包围盒的功能,通过对AutoCAD软件的二次开发,利用程序自动确定出任意形状物体的最小体积包围盒或最小面积包围盒,以适应于不同的应用场合。     

基于Soft-NMS的候选框去冗余加速器设计

目标检测任务通常使用非极大值抑制算法(NMS)删除卷积神经网络输出的冗余候选框.Soft-NMS使用逐步衰减候选框得分值的方法代替Hard-NMS中直接删除大于预定义阈值候选框的方法,可以避免误删图像中重叠的目标候选框,提高目标检测任务的准确率.但是,频繁地改变候选框得分值使得Soft-NMS较Hard-NMS更为复杂...     

Relative importance of expertise, lifting height and weight lifted on posture and lumbar external loading during a transfer task in manual material handling

The objective of this study was to measure the effect size of three important factors in manual material handling, namely expertise, lifting height and weight lifted. The effect of expertise was evaluated by contrasting 15 expert and 15 novice handlers, the effect of the weight lifted with a 15-kg box and a 23-kg box and the effect of lifting height with two different box heights: ground level and a 32 cm height. The task consisted of transferring a series of boxes from a conveyor to a hand trolley. Lifting height and weight lifted had more effect size than expertise on external back loading variables (moments) while expertise had low impact. On the other hand, expertise showed a significant effect of posture variables on the lumbar spine and knees. All three factors are important, but for a reduction of external back loading, the focus should be on the lifting height and weight lifted. PRACTITIONER SUMMARY: The objective was to measure the effect size of three important factors in a transfer of boxes from a conveyor to a hand trolley. Lifting height and weight lifted had more effect size than expertise on external back loading variables but expertise was a major determinant in back posture.