等速技术检测膝关节伪装瘫痪的可靠性

目的评估等速技术对健康青年志愿者配合及伪装膝关节运动的重测可靠性。方法 30名健康青年志愿者在30°/s角速度下间隔45min执行2次向心全力配合及半力、随机力量伪装膝关节屈伸运动,对2次测试的峰力矩(peak torque,PT)值进行统计学分析。结果全力测试的相对可靠性ICC值(0.892~0.956)大于半力和随机力量的ICC值(0.246~0.481),并且全力的绝对可靠性各参数数值均小于半力和随机力量。结论在配合条件下,等速技术对健康青年膝关节PT的检测具有良好的可靠性,并高于伪装条件下的可靠性。     

正常人不同姿势膝关节等速肌力测试

目的：研究应用CON-TREX Biomechanical Test and Training系统测试膝关节肌力时,分析不同姿势对测试结果的影响,从而选择最适用于法医学鉴定的肌力检测条件。方法对52例正常人分别在未固定姿势和固定姿势两种条件下进行60°/s和30°/s速度的等速肌力测试,对这两种条件下双侧膝关节屈、伸肌的峰力矩（peak torque,PT）、峰力矩角度（peak torque angle,PTA）进行统计学分析。结果在未固定姿势,两种测试速度下受试者膝关节屈、伸肌的PT双侧比较差异均存在统计学意义（P〈0.05）;在固定姿势,两种测试速度下受试者膝关节屈、伸肌的PT双侧比较差异均无统计学意义（P〉0.05）。不论受试者身体姿势是否加以固定,其膝关节屈、伸过程的PTA双侧比较差异均无统计学意义（P〉0.05）。结论受试者姿势的规范对测试结果,对PT有影响,因此,在进行膝关节等速肌力测试时应规范受试者的姿势。     

等速肌力检测在辨别伪肢体瘫中的应用

目的：探讨等速肌力检测在法医学鉴定中辨别伪肢体瘫的应用价值,为建立规范、统一的法医学肌力评价方法提供科学依据。方法选取骨折或神经损伤者57例（损伤组）和伪装者128例（伪装组）,应用等速肌力测试系统对损伤组两侧膝关节与伪装组膝关节进行检测,对双侧膝关节屈、伸肌的峰力矩（peak torque,PT）、峰力矩角度（peak torque angle,PTA）进行比较,归纳两组膝关节的力矩-时间图特征。结果损伤组膝关节屈、伸肌的双侧PT的比较差异有统计学意义（P＜0.05）,PTA的比较差异无统计学意义（P＞0.05）;伪装组膝关节屈、伸肌的双侧PT、PTA的比较差异均有统计学意义（P＜0.05）。损伤组两侧膝关节力矩-时间图主要为单峰前置型,伪装组膝关节力矩-时间图主要为多峰型。结论在法医学鉴定中,等速肌力检测在辨别伪肢体瘫方面可以通过力矩-时间图的相关特点进行评价。     

基于深度学习实现成人坐骨耻骨支内侧缘的性别推断

目的 探究深度学习技术在中国汉族人群CT三维重建图像自动性别识别中的可靠性和准确率。方法 收集20～85岁汉族人群骨盆CT影像学数据700例（男性350例，女性350例），将其重建为三维虚拟骨骼模型，并截取坐骨耻骨支内侧缘（medial aspect of the ischiopubic ramus,MIPR）特征区域图像。采用Inception v4作为图像识别模型，以初始化学习和迁移学习两种方式进行训练。随机选取80%的图像作为训练验证集，20%的图像作为测试集。将左右两侧MIPR图像进行单独以及合并训练。之后使用总准确率、女性准确率、男性准确率等指标进行模型的性能评价。结果将左右两侧MIPR图像单独进行初始化学习训练，右侧MIPR模型的总准确率为95.7%，其中女性准确率为95.7%、男性准确率为95.7%；左侧MIPR模型的总准确率为92.1%，其中女性准确率为88.6%、男性准确率为95.7%。将左右两侧MIPR图像合并以初始化学习进行训练，模型的总准确率为94.6%，其中女性准确率为92.1%、男性准确率为97.1%。将左右两侧MIPR图像合并以迁移学习进行训练，模型的总准...     

基于深度学习的颅脑损伤机制自动化鉴别

目的 应用卷积神经网络Inception＿v3模型进行基于颅脑CT图像的加、减速性损伤自动鉴别,探讨深度学习技术在法医学颅脑损伤成伤机制推断中的应用前景。方法 收集190例加、减速性脑损伤的影像学资料作为实验组,另选取130例正常颅脑的影像学资料作为对照。将上述320例影像学资料根据随机抽样的方法分为训练验证集和测试集。采用准确率、精确率、召回率、F1值及AUC值评估模型分类性能。结果 模型在训练过程和验证过程中对3种图像（加速性损伤、减速性损伤及正常颅脑）分类的最高准确率分别为99.00%、87.21%,满足要求。使用优化后的模型对测试集数据进行测试,该模型在测试集中的三分类准确率为87.18%,识别加速性损伤的精确率、召回率、F1值及AUC值分别为84.38%、90.00%、87.10%、0.98,识别减速性损伤的各值分别为86.67%、72.22%、78.79%、0.92,识别正常颅脑的各值分别为88.57%、89.86%、89.21%、0.93。结论 Inception＿v3模型在基于颅脑CT图像区分加、减速性损伤中具有应用潜力,有望成为推断头部致伤方式的一种辅助工具。     

基于YOLOv3的硅藻电子显微镜图像识别模型的构建及应用

目的 基于YOLOv3模型构建硅藻电子显微镜图像识别模型,测试在实际案例中的应用效果,探讨该模型在法医学硅藻鉴定中的优势。方法 选取25 000张1 500×的硅藻电子显微镜图像作为初始图像集,经专家标注和图像处理后,输入YOLOv3网络训练识别模型。收集8例溺死案例肺、肝、肾组织样本的硅藻电子显微镜图像,分别在0.4、0.6、0.8的阈值下利用该模型识别硅藻,同时进行人工识别,以识别速度、召回率和查准率评价该模型的实际应用效果。结果 模型在验证集和测试集上的平均精度均值分别为94.8%和94.3%,平均召回率分别为81.2%和81.5%。实际应用中,该模型识别速度是人工识别速度的9倍以上。阈值为0.4时,肺组织硅藻总体召回率均值达89.6%、总体查准率均值达87.8%;肝、肾组织硅藻总体召回率达100%、查准率低于5%。随着阈值的增加,各组织硅藻召回率下降、查准率上升;模型在肺组织的F1分数随着阈值升高而降低,在肝、肾组织的F1分数随着阈值升高而升高。结论 基于YOLOv3的硅藻电子显微镜图像识别模型能够快速识别硅藻图像,并能在一定的阈值下取得良好的硅藻召回率和肺组织较高的硅藻查准率...     

4种深度学习图像分类算法在人工智能硅藻检验中的比较

目的 选择深度学习图像分类算法中复杂性和准确性较为平衡的4种算法进行硅藻的自动识别,探究最适用于硅藻识别的分类算法,为法医学自动化硅藻检验研究提供数据参考。方法 建立真实水中尸体肺组织消化液涂片的“硅藻”“背景”小样本量数据集（20 000张）,用于4种算法（VGG16、ResNet50、InceptionV3和Inception-ResNet-V2）模型的训练、验证和测试。绘制受试者工作特征曲线、混淆矩阵并计算召回率、查准率、特异性、准确率及F1分数,对各模型性能进行系统性评估。结果 InceptionV3的硅藻识别性能明显优于其他3种算法,具有更为均衡的硅藻查全（89.80%）与查准（92.58%）性能;VGG16和InceptionResNet-V2的硅藻识别性能相当,虽无法做到硅藻查全与查准的性能均衡,但其识别能力尚可接受;特征,且都以硅藻区域为主要识别依据。结论 包含Inception结构的模型,在硅藻特征提取方面具有更强的指向性和靶向性。其中,InceptionV3算法能够更为准确、靶向地提取到硅藻特征,具有最优的硅藻识别性能,更适合应用于日常法医学硅藻检验。     

基于数据挖掘算法的骨盆损伤判别模型

目的 利用主成分分析、偏最小二乘法对基于骨盆CT图像提取的特性信息进行降维,利用降维的数据建立判别骨盆是否受伤的支持向量机分类判别模型,评估其应用的可行性。方法 将采集的正常和受伤骨盆CT图像146例分别随机提取80%作为训练集,用于模型拟合;剩余20%作为测试集,用于模型准确性的检验。通过CT图像输入、图像预处理、特征提取、特征降维、特征选择、参数选择、模型建立和模型比较等步骤,建立骨盆是否受伤的判别模型。结果 偏最小二乘法降维方法优于主成分分析降维方法,支持向量机模型优于朴素贝叶斯模型。基于12个偏最小二乘因子建立的骨盆是否受伤支持向量机分类判别模型的建模集、留一法交叉验证和测试集结果准确率分别为100%、100%和93.33%。结论 基于CT图像建立的骨盆是否受伤数据挖掘模型在评估骨盆损伤中具有比较高的准确性,为骨盆损伤的自动快速识别奠定基础。     

基于多尺度特征融合的穿鞋足迹识别网络

近年来，随着人工智能的快速发展，赤足足迹和穿袜足迹的自动识别均取得了不错的效果，但因为需要脱鞋，应用场景受到了很大的限制。单枚穿鞋足迹识别由于鞋底花纹的多样性、随机性给智能识别造成了很大的阻碍，识别准确率普遍较低，成为了一个具有挑战性的任务。本文聚焦于不同人穿同一种鞋的足迹识别，提出一种基于多尺度特征融合的穿鞋足迹识别网络。采集大量的足迹样本，通过旋转、平移、加噪声等方式模拟犯罪现场足迹图像可能产生的变化，制作了足迹数据集，然后以ResNet为骨干网络，利用双向金字塔特征融合模块充分融合足迹图像的深层特征和浅层特征，最后针对鞋底花纹变化造成网络识别准确率降低的问题，提出了用迁移学习解决的思路，让网络学习未知花纹与现有花纹之间的联系，快速拟合出模型。经过训练测试，本文提出的方法在自制足迹数据集上识别准确率达到了93.1%,CMC评价指标也明显优于其他网络。在新的足底花纹上，迁移学习对比从头开始训练，速度更快、准确率更高。大量的实验证明，本文提出的穿鞋足迹识别网络识别准确率更高，提出的迁移学习的方法在面对新的鞋底花纹时，能够实现更好的效果。     

基于深度学习对口腔全景摄影片进行性别推断的研究

目的 探究深度学习技术对中国汉族人群口腔全景摄影图像（OPG）进行性别推断的可行性和准确性。方法 收集18～70岁汉族人群OPG影像数据10 600份（男性5 300份,女性5 300份）,按照8:1:1比例随机划分为训练集、验证集、测试集。采用MobileNetV2、Swin Transformer模型进行训练,通过准确率、F1分数及GradCAM算法评估模型分类性能及可视化展示。结果 MobileNetV2、Swin Transformer Small和Swin Transformer Tiny模型的准确率为97.57%、95.13%、96.28%,其中MobileNetV2模型表现最佳。Grad-CAM算法显示男性OPG图像主要关注左右侧下颌支及牙槽骨,而女性的OPG图像主要关注上颌窦、左侧下颌支及左侧后牙槽骨。结论 本研究基于深度学习技术构建的中国汉族人群OPG图像性别推断模型具有较高的准确性及泛化能力,为法医学性别推断提供了新的方法。     

“AGING OUT: DON'T MISS OUT”—A MODEL OF COMMUNITY LEGAL EDUCATION

Typically, community legal education projects are ad hoc efforts where attorneys, law students, paralegals, or community organizers present how‐to workshops, develop manuals, or lead informational or training sessions for other attorneys, advocates, or members of the community. However, community education efforts can also promote and encourage public awareness, community engagement, and advocacy. This essay explores why community legal education efforts are needed and particularly well suited for youth transitioning out of foster care, and examines a specific community legal education effort that is being implemented with and on behalf of foster youth transitioning out of New Jersey's foster care system. The hope is that by studying this example, others may learn from it and be inspired to replicate it, as part of a state's or a community's overall effort to aid, empower, and engage foster youth.     

Dealing with Wayward Desire

The exercise of synchronic self-control is the way in which an actor can attempt to bring a desire into alignment with his better judgement at the moment and during the interval of time over which, but for the exercise of such self-control, the desire would become the actor’s preponderant desire, which the actor would then translate into an act contrary to his better judgment. The moral psychology of an actor who fails to achieve such self-control can be analyzed in two ways. One way is meant to be consistent with compatibilist metaphysics; the other with libertarian metaphysics. The implications of these analyses for the criminal law are complicated, but perhaps the most important is this: the criminal law should in principle recognize a partial excuse for an actor who exercises synchronic self-control but who gives up his effort because he believes that he can no longer continue to resist. His effort to achieve self-control thus fails, and he ends up translating into action the very desire he set out to control.

"Sentinel effect" shows fraud control effort works

The author analyzes the fraud and abuse track record of his agency in the past years, and discuss why that effort has had beneficial results for the healthcare industry and the federal fisc. The benefits of this enforcement effort flow directly from the specific enforcement actions, and indirectly from the response of others in the industry to the standards set forth in those actions. Mr. Thornton disputes assertions of some in the industry that the enforcement efforts are unfair, arbitrary, or draconian. Rather, they are an appropriate response to a documented problem with which the government is justifiably concerned.     

中国汉族女性青少年法医学活体骨龄推断数学模型的建立

目的建立推断中国汉族女性青少年活体骨龄的数学模型。方法摄取华中、华南及华东等地区的838名年龄介于11～20周岁正常女性青少年双侧锁骨胸骨端以及左侧肩、肘、腕、髋、膝、踝关节的X线片。依据青少年骨发育分级标准对24项骨骼发育指标进行阅片、分级,结合考虑身高、体质量及地区等影响因素．应用SAS8．1及SPSS11．0软件进行统计学处理,探索各指标与年龄的相关性。结果建立了我国汉族女性青少年利用锁骨胸骨端及6大关节骨骺闭合程度联合推断活体年龄的多元回归数学模型．推导出判定我国汉族女性青少年是否已满14、16和18周岁的Fisher’S两类判别分析方程。结论本研究所建立的判定活体年龄的数学模型丰富了活体年龄的法医学鉴定方法,有利于提高活体骨龄鉴定方法的科学性和结论的准确性。     

The effects of dosed tobacco in evidentiary breath testing using non-drinking subjects

A total of seventeen subjects were administered breath tests with alcohol dosed tobacco to see if there was an interference with the evidentiary breath testing. Fourteen subjects provided one set of two breath samples without the dosed tobacco followed by a set of two breath samples with the dosed tobacco. The other three subjects provided one breath sample without the dosed tobacco and then one breath sample with the dosed tobacco within the same testing sequence. Eight subjects had breath test readings of 0.00g/210L with the dosed tobacco. Mouth alcohol was detected with the dosed tobacco in six of the subjects, and a reading of 0.01g/210L, 0.04g/210L, and 0.05g/210L were found in five of the subjects. If the officer follows the directive of checking the mouth for a foreign substance and following a 15-20min observation/deprivation period, a false positive result will likely be avoided. If the officer does not find tobacco when checking the mouth for a foreign substance, and dosed tobacco is present during the breath test, most likely there would not be a measurable amount of alcohol to report or there would be a mouth alcohol reading from the sample.     

Results of Colombian exercise interlaboratory quality control 2012

Participation in interlaboratory quality control exercises is one of the main mechanisms currently used for quality assurance and continuous improvement of the trials. The objective of this study was to design, to manage and to evaluate the Colombian Exercise Interlaboratory Quality Control 2012 (CEIQC-2012). The CEIQC-2012 included both practical and theoretical exercises. For practical exercise three samples were provided, two from blood and one from buccal swab, all on FTA cards, participants were requested to process the samples according to the methods and the markers used routinely in their own laboratories. For theoretical section four exercises were sent, and only one was mandatory, the remain have different degrees of difficulty and were optional. In the mandatory exercise, the participants were asked to calculate the partial and total IP of 15 autosomal STRs markers of an alleged father and a son. This exercise involved 28 laboratories from 6 Latin American and Caribbean countries (Brazil, Ecuador, Peru, Panama, Dominican Republic and Colombia), all reported results for the theoretical mandatory and 27 for the practical. Fifty-four STR markers distributed in autosomal, Y and X chromosomes were under consensus. The Proficiency Test conducted through the Colombian National Reference Laboratory has become a useful tool for quality assurance of all Colombian laboratories and some of Latin America and Caribbean that perform DNA testing to establish biological relationships. This exercise is also an excellent opportunity for constant experts training in the region.     

Effect of biofeedback on the detection of deception

This study examined the effect of audio electrodermal biofeedback training on the detection of deception. The subjects consisted of 68 volunteers enrolled in selected undergraduate college courses. Each subject was required to commit a mock murder, after which a polygraph examiner administered a series of five consecutive "lie detector" tests to ascertain the facts involved in his/her murder. Before testing, subjects were randomly assigned to either a biofeedback condition or to a control group. The detection efficiency associated with the subject's respiration responses was significantly enhanced by simultaneous auditory biofeedback given during the polygraph testing; however, the feedback's effect upon the detection rates associated with the electrodermal measures that it was reflecting was neither statistically significant nor in the same direction. The results support the premise that audio biofeedback might be useful in enhancing respiration's detection efficiency during polygraph testing.