基于图像显著性的人工视觉图像处理策略

目的:在假体设备中的视觉信息处理模块引入适当的图像处理策略,优化低分辨率下的人工视觉信息,是解决假体植入者获取的视觉信息有限问题的一种可行方法。方法:基于全局亮度对比度特征的图像显著性检测算法（LC）,结合颜色空间变换和视觉注意力仿真处理模型,提出一种面向视网膜假体人工视觉信息处理策略。通过两个标准图像测试数据集对图像处理结果进行评估,同时对流行病学调查结果选取的盲人常用物体图片进行仿真。结果:评估结果验证了在提取图像前景方面与原始LC算法对比的优越性,仿真结果验证了在人工视觉条件下应用的可行性。结论:本文提出的策略有助于视网膜假体植入者在日常生活场景中更好地完成物体识别等基本视觉任务。     

基于实例分割和显著性计算的人工视觉多目标优化处理

通过实例分割Swin-Transformer提取分割所有前景对象,融合亮度、大小和位置图像显著性特征,提出模拟人类视觉注意机制的多特征融合注意力层级计算模型,为不同级别的前景物体采用适合的光幻视分辨率和亮度表达,实现不同的刺激编码策略进行层级优化处理。通过人工假体视觉的仿真试验表明,在所提出的多目标层级优化表达策略下,试验被试完成多目标识别的准确率、识别时间表现具有一定的显著提升。利用深度学习实例分割技术,层级化光幻视编码以仿生人类视觉选择性注意,达到增强假体植入者在复杂场景下的多物体感知,为视觉假体图像信息编码和优化处理研究的发展与应用提供参考。     

仿真假体视觉下的人脸识别研究进展

目的：综述基于仿真假体视觉的人脸识别研究的主要进展。方法：回顾了近年来各研究小组基于仿真假体视觉的人脸识别研究进展,讨论假体视觉下人脸识别的可能性,分析了假体视觉下人脸识别的影响因素。结果：在仿真假体视觉下,分辨率和缺失率是影响人脸识别的主要因素,分辨率增加或是缺失率减小,对人脸的识别率有显著提高;其它光幻视参数中,一定程度内的灰度提高对假体下的人脸识别率有显著提高,而点大小、点间隙、对比度对其影响稍小,正方形的光幻视阵列排布的识别率高于六边形和极坐标排布。另外,图像处理策略如基于感兴趣区域的放大和对比度增强等,能帮助被试进行人脸识别。结论：由心理物理学方法得出,在假体视觉下,仍能完成基本的人脸识别任务;改变光幻视的参数或图像处理策略,会影响识别率。这些结果将帮助研究者们优化视觉假体中的信息处理和图像处理策略,并为术后康复训练提供了实验理论依据。     

一种视觉假体的视觉显著图提取新方法

目的 一幅图像或场景的显著性区域代表它们的主要内容（显著目标）.由于视觉假体可植入电极的数量有限,只有低分辨率的图像对其才有用,因此提取显著性区域有助于视觉假体捕捉到场景中的显著目标.方法 Itti模型是一个显著性检测模型,它检测到的显著性区域与人的视觉感知有差异,显著目标的边界不明确.笔者去除了Itti模型提取的方向和色彩特征,将红（R）、绿（G）、蓝（B）三基色（RGB）图像转换到对应于HSI颜色空间上的色调（H）、纯度（S）、亮度（I）3个新特征分量,对hti模型进行优化改进.在显著图中,将落在显著目标内的显著点面积与总显著点面积的比值定义为显著图精确度;以显著图精确度为提取显著图方法的测度,对改进前后2种方法进行比较.结果 利用改进方法提取的显著图比Itti模型显著图精确度提高了约20％;在检测显著性区域时所用时间减少近50％.结论 提出了一种用于人工视觉的获取显著目标的方法,本算法可以得到更加精确的显著性结果,且可缩短运行时间.     

合成孔径序列波束形成轴向运动补偿方法

合成孔径序列波束方法是一种新颖的医学超声成像方法,采用两个阶段的波束形成,在传统的超声成像系统中实现合成孔径成像,在不需要存储和传输大量射频回波数据的情况下,提高医学超声图像的分辨率。该方法的前提是假设成像目标静止不动,而通过仿真分析发现运动会造成成像目标位置错误。针对此问题,提出一种合成孔径序列波束形成运动估计和补偿方法：首先通过在同一位置连续发射接收两次以获取用于运动估计的数据,然后采用互相关方法对第一阶段波束形成得到的低分辨率图像进行运动估计和补偿,再对其做第二次波束形成得到高分辨率图像。Field II仿真结果显示,所提出的运动补偿方法可以得到正确的目标位置。对于运动速度为0.1 m/s的点目标,运动补偿后的平均横向分辨率与静止点目标相比仅降低2.03%,对比度降低2.7 dB。对于运动速度为0.2 m/s的囊目标,运动补偿后图像的对比度分辨率较静止情况仅降低8.53%,进一步说明所提出的运动补偿方法有效。     

新型运动模式铰链膝关节假体磨损有限元分析

目的 对新型球轴运动模式铰链膝关节假体进行有限元分析,研究假体磨损仿真的方法及运动方式对假体磨损的影响。方法 基于球轴假体接触力学有限元模型,根据Archard磨损理论,建立磨损有限元模型,以各种运动数据为加载条件,模拟生理活动下膝关节的力学环境,研究球轴假体磨损情况。结果 对于胫骨衬垫,上下楼时平均和最大接触应力均高于步行,上楼时累积磨损体积大于下楼和步行,且磨损均主要发生在衬垫下表面。对于球轴衬套,仅步行时存在短时间接触和磨损,累积磨损体积为0.19 mm³。结论 铰链膝关节假体的球轴运动模式可改善置换后膝关节力学环境,降低衬套磨损,延长假体生存期。有限元仿真可有效预测铰链假体的磨损,为其设计及改进提供理论支持。     

基于有限元方法的全髋关节假体个体化选型分析

目的 研究ISO 7206标准对全髋关节置换术临床应用的指导意义。方法 建立肌骨数值模型,对正常行走步态进行仿真,以获得下肢的运动学和动力学参数;建立对应的全髋关节假体有限元模型,应用步态载荷进行计算,并对比ISO标准的有限元模型计算结果。结果 正常行走步态下,髋关节力分别在20%和54%步态周期出现峰值,以此作为有限元计算的步态载荷,得到20%步态周期时假体应力最大;松动模型中假体柄上最大应力大于无松动模型中假体柄上最大应力,且应力分布趋势存在差异;分析对比ISO测试和不同体重人体步态载荷下的假体最大应力,得到ISO测试中最大应力水平对应108~142 kg体重载荷下的假体最大应力。结论ISO测试中合格的假体可满足100 kg体重人体正常步态下的强度要求。     

抑制MR成像中刚体运动伪影的新方法

目的从成像空间出发,提出了一种新的纠正MRI刚性平移运动伪影的方法.方法利用梯度读出方向和相位编码方向的运动特性不同,分别采用不同的方法来消除刚性运动伪影.首先读出方向的运动,通过追踪频谱边缘非零区域和零区域的偏差进行估计,然后在频谱反方向移动相同的量来消除;利用改进的Snake算法,即梯度向量场方法提取目标区域的边界,然后利用相位迭代恢复算法消除残留的读出方向亚像素级伪影和相位编码方向的伪影.结果按国际通用方法生成SL头颅模板,通过对模板的仿真试验,证明修正后图像的信噪比大大提高,验证了方法的有效性和可靠性.结论本文提出的方法能够有效消除MR图像平移运动伪影,与传统的相位迭代恢复算法相比,对于较大运动伪影修正效果更好.     

体感电刺激诱发事件相关电位的研究

目的对视觉、听觉、体感3种不同模态下靶刺激诱发的事件相关电位(ERP)进行比较研究,探讨体感电刺激作为脑机接口(BCI)一种新的信号诱发模式的可能性,为基于体感ERP的BCI研究提供理论依据。方法选择17例视力或矫正视力正常、听力正常、躯体感觉正常且无任何大脑病史的被试者,其中男性8例,女性9例;年龄20~26岁,平均年龄22.6岁;均为右利手。分别记录17例健康的被试者在视觉、听觉、体感单通道靶刺激下诱发的脑电图;对3类靶刺激下ERP的时域参数(幅值、潜伏期)、行为学数据(反应时间、错误率)、脑源定位进行比较分析。结果 3类靶刺激模式下的ERP波形具有相似性,体感电刺激诱发的ERP幅值与视觉、听觉靶刺激相比无显著性差异;体感电刺激诱发ERP的峰值潜伏期显著长于视觉靶刺激;体感电刺激的反应时间显著长于视觉靶刺激,错误率也高于视觉、听觉靶刺激;体感电刺激诱发ERP的脑内源与视觉靶刺激相比具有相似性。结论相比于视觉、听觉靶刺激,大脑对于体感电刺激的探测难度高,敏感程度低;但从ERP的波形和幅值上看,体感电刺激可以诱发出稳定的、可被检测到的ERP波形,完全有可能应用于BCI系统作为一种新的ERP诱发模式。     

初次全髋关节置换中两种直径股骨头活动度的对比

背景：人工全髋关节置换后脱位是置换后最常见的近期并发症,引起脱位的因素很多。在术者的可控制的因素中,股骨头直径可能是影响脱位发生率的因素之一。一般认为较大的假体头更符合髋关节生物力学特点,能够增加关节稳定性。 目的：观察初次全髋关节置换术中不同股骨头直径对髋关节活动度的影响,以期为临床选择理想直径的股骨头假体提供依据。 方法：收集乌兰察布市中心医院骨科自2009年8月至2012年8月采用28 mm及32 mm直径的股骨头假体行初次全髋置换77例(87髋)患者的病历资料。全部髋关节假体由ZIMMER公司提供,髋关节活动界面为聚乙烯对金属。股骨头假体采用28 mm直径51髋,32 mm直径36髋。手术入路均为后侧入路。术中测髋关节屈髋90°时髋关节内、外旋至脱位的活动度,并在不同直径股骨头假体组间进行对比分析。 结果与结论：全部患者术中未更改术式。将两种直径股骨头组间脱位时度数总和进行统计学分析,发现差异无显著性意义(P > 0.05)。其中股骨颈骨折患者行人工全髋关节置换术中,两种不同直径人工股骨头假体的术中活动度差异无显著性意义(P > 0.05);股骨头缺血性坏死患者行人工全髋关节置换术中,两种不同直径人工股骨头假体的术中活动度差异无显著性意义(P > 0.05)。提示与28 mm股骨头相比,32 mm股骨头未能增加全髋关节置换术中活动度,置换后随访2年两种直径的人工股骨头对于全髋关节置换后脱位无影响。中国组织工程研究杂志出版内容重点：人工关节;骨植入物;脊柱;骨折;内固定;数字化骨科;组织工程      

Contrast and assimilation in motion perception and smooth pursuit eye movements

The analysis of visual motion serves many different functions ranging from object motion perception to the control of self-motion. The perception of visual motion and the oculomotor tracking of a moving object are known to be closely related and are assumed to be controlled by shared brain areas. We compared perceived velocity and the velocity of smooth pursuit eye movements in human observers in a paradigm that required the segmentation of target object motion from context motion. In each trial, a pursuit target and a visual context were independently perturbed simultaneously to briefly increase or decrease in speed. Observers had to accurately track the target and estimate target speed during the perturbation interval. Here we show that the same motion signals are processed in fundamentally different ways for perception and steady-state smooth pursuit eye movements. For the computation of perceived velocity, motion of the context was subtracted from target motion (motion contrast), whereas pursuit velocity was determined by the motion average (motion assimilation). We conclude that the human motion system uses these computations to optimally accomplish different functions: image segmentation for object motion perception and velocity estimation for the control of smooth pursuit eye movements.     

From following edges to pursuing objects

Primates can generate accurate, smooth eye-movement responses to moving target objects of arbitrary shape and size, even in the presence of complex backgrounds and/or the extraneous motion of non-target objects. Most previous studies of pursuit have simply used a spot moving over a featureless background as the target and have thus neglected critical issues associated with the general problem of recovering object motion. Visual psychophysicists and theoreticians have shown that, for arbitrary objects with multiple features at multiple orientations, object-motion estimation for perception is a complex, multi-staged, time-consuming process. To examine the temporal evolution of the motion signal driving pursuit, we recorded the tracking eye movements of human observers to moving line-figure diamonds. We found that pursuit is initially biased in the direction of the vector average of the motions of the diamond's line segments and gradually converges to the true object-motion direction with a time constant of approximately 90 ms. Furthermore, transient blanking of the target during steady-state pursuit induces a decrease in tracking speed, which, unlike pursuit initiation, is subsequently corrected without an initial direction bias. These results are inconsistent with current models in which pursuit is driven by retinal-slip error correction. They demonstrate that pursuit models must be revised to include a more complete visual afferent pathway, which computes, and to some extent latches on to, an accurate estimate of object direction over the first hundred milliseconds or so of motion.     

Human updating of visual motion direction during head rotations

Previous studies have demonstrated that human subjects update the location of visual targets for saccades after head and body movements and in the absence of visual feedback. This phenomenon is known as spatial updating. Here we investigated whether a similar mechanism exists for the perception of motion direction. We recorded eye positions in three dimensions and behavioral responses in seven subjects during a motion task in two different conditions: when the subject's head remained stationary and when subjects rotated their heads around an anteroposterior axis (head tilt). We demonstrated that after head-tilt subjects updated the direction of saccades made in the perceived stimulus direction (direction of motion updating), the amount of updating varied across subjects and stimulus directions, the amount of motion direction updating was highly correlated with the amount of spatial updating during a memory-guided saccade task, subjects updated the stimulus direction during a two-alternative forced-choice direction discrimination task in the absence of saccadic eye movements (perceptual updating), perceptual updating was more accurate than motion direction updating involving saccades, and subjects updated motion direction similarly during active and passive head rotation. These results demonstrate the existence of an updating mechanism for the perception of motion direction in the human brain that operates during active and passive head rotations and that resembles the one of spatial updating. Such a mechanism operates during different tasks involving different motor and perceptual skills (saccade and motion direction discrimination) with different degrees of accuracy.     

Vection increases the magnitude and accuracy of visually evoked postural responses

Movement of large visual scenes induces an illusion of self-motion (vection) and postural responses. We investigated if the conscious perception of self-motion influences the magnitude and directional accuracy of visually evoked postural responses. Five normal subjects fixated the centre of a large disk rotating in the roll (coronal) plane. The disk was placed either in front of the subjects or obliquely 30 deg to their right or left; in these oblique positions disk fixation was achieved by horizontal ocular deviation alone (i.e. no neck deviation). Subjects indicated their subjective perceptual status, either vection or object motion, with a push button. The results confirmed that the direction of the visually evoked postural response was reoriented according to the different eye-disk positions. In addition, both the magnitude of the postural response and the accuracy of its alignment with the disk rotational plane were significantly increased during vection periods. The results show that conscious perception of self-motion enhances visuopostural performance. Since conscious perception is likely to arise at cortical levels, the findings indicate that the cortex is one of the sites where gaze direction interacts with retinal motion signals to provide a self-motion signal in body-centric co-ordinates. Such interaction provides a substrate for spatial representation during motion in the environment. Electronic Publication     

Collision judgment of objects approaching the head

Recent investigations have indicated that human perception of the trajectory of objects approaching in the horizontal plane is precise but biased away from straight ahead. This is remarkable because it could mean that subjects perceive objects that approach on a collision course as missing the head. Approach within the horizontal plane through the eyes and the fixation point (the plane of regard) is special, as general motions will also have a component of motion perpendicular to the plane of regard. Thus, we investigated three-dimensional motion perception in the vicinity of the head, including vertical components. Subjects judged whether an object that moved in the mid-sagittal plane was going to hit below or above a well-known reference point on the face like the center of the chin or the forehead (perceptual task). Tactile and proprioceptive information about the reference point significantly improved precision. Precision did not change with distance of the approaching target or with fixation direction. Bias was virtually absent for these vertical motions. When subjects pointed with their index finger to the perceived location of impact on their face (visuo-motor task), they overestimated (1.7 cm) the horizontal eccentricity of the point of impact (pointing task). Vertical bias, however, was again virtually absent. Interestingly, when trajectories intersected the plane of regard, higher precision was observed in the perceptual task regardless of the other conditions. In contrast, neither bias nor precision of the pointing task changed significantly when the trajectories intersected the plane of regard. When asked to point to the location where a trajectory intersected the plane of regard, subjects overestimated the depth component of this intersection location by about 3 cm. The absence of perceptual and pointing bias in the vertical direction in contrast to the clear horizontal bias suggests that different (combinations of) cues are used to judge these components of the trajectory of an approaching object. The results of our perceptual task suggest a role for somatosensory signals in the visual judgment of impending impact.     

Smooth-pursuit eye movements elicited by first-order and second-order motion

 The perception of the displacement of luminance-defined contours (i.e., first-order motion) is an important and well-examined function of the visual system. It can be explained, for example, by the operation of elementary motion detectors (EMDs), which cross-correlate the spatiotemporal luminance distribution. More recent studies using second-order motion stimuli, i.e., shifts of the distribution of features such as contrast, texture, flicker, or motion, extended classic concepts of motion perception by including nonlinear or hierarchical processing in the EMD. Smooth-pursuit eye movements can be used as a direct behavioral probe for motion processing. The ability of the visual system to extract motion signals from the spatiotemporal changes of the retinal image can be addressed by analyzing the elicited eye movements. We measured the eye movement response to moving objects defined by two different types of first-order motion and two different types of second-order motion. Our results clearly showed that the direction of smooth-pursuit eye movements was always determined by the direction of object motion. In particular, in the case of second-order motion stimuli, smooth-pursuit did not follow the retinal image motion. The latency of the initial saccades during pursuit of second-order stimuli was slightly but significantly increased, compared with the latency of saccades elicited by first-order motion. The processing of second-order motion in the peripheral visual field was less exact than the processing of first-order motion in the peripheral field. Steady state smooth-pursuit eye speed did not reflect the velocity of second-order motion as precisely as that of first-order motion, and the resulting retinal error was compensated by saccades. Interestingly, for slow second-order stimuli we observed that the eye could move faster than the target, leading to small, corrective saccades in the opposite direction to the ongoing smooth-pursuit eye movement. We conclude from our results that both visual perception and the control of smooth-pursuit eye movements have access to processing mechanisms extracting first- and second-order motion. Received: 26 August 1996 / Accepted: 8 November 1996     

Hitting moving targets

Previous work has indicated that people do not use their judgment of a target's speed to determine where to hit it. Instead, they use their judgment of the target's changing position and an expected speed (based on the speed of previous targets). In the present study we investigate whether people also ignore the target's apparent direction of motion, and use the target's changing position and an expected direction of motion instead. Subjects hit targets that moved in slightly different directions across a screen. Sometimes the targets disappeared after 150 ms, long before the subjects could reach the screen. This prevented subjects from using the target's changing position to adjust their movements, making it possible to evaluate whether subjects were relying on the perceived or an expected (average) direction to guide their movements. The background moved perpendicular to the average direction of motion in some trials. This influences the target's perceived direction of motion while leaving its perceived position unaffected. When the background was stationary, subjects hit disappearing targets along their trajectory, just as they hit ones that remained visible. Moving the background affected the direction in which subjects started to move their hand, in accordance with the illusory change in direction of target motion. If the target disappeared, this resulted in a hit that was systematically off the target's trajectory. If the target remained visible, subjects corrected their initial error. Presumably they did so on the basis of information about the target's changing position, because if the target disappeared they did not correct the error. We conclude that people do use the target's perceived direction of motion to determine where to hit it. Thus the perceived direction of motion is treated differently than the perceived speed. This suggests that the motion of an object is not broken down into speed components in different directions, but that speed and direction are perceived and used separately.     

Using visual direction in three-dimensional motion perception

The eyes receive slightly different views of the world, and the differences between their images (binocular disparity) are used to see depth. Several authors have suggested how the brain could exploit this information for three-dimensional (3D) motion perception, but here we consider a simpler strategy. Visual direction is the angle between the direction of an object and the direction that an observer faces. Here we describe human behavioral experiments in which observers use visual direction, rather than binocular information, to estimate an object's 3D motion even though this causes them to make systematic errors. This suggests that recent models of binocular 3D motion perception may not reflect the strategies that human observers actually use.     

An unbiased measure of the contributions of chroma and luminance to saccadic suppression of displacement

Our previous study revealed that alginate gel cross-linked with covalent bonds promoted peripheral nerve regeneration in the cat and rat. The present study analyzed nerve regeneration through alginate gel in the early stages within 2 weeks and the late stages up to 21 months after implantation. Four days after surgery, regenerating axons grew without Schwann cell investment through the partially degraded alginate gel, being in direct contact with the alginate without a basal lamina covering. Numerous mast cells infiltrated into the alginate. One to 2 weeks after surgery, regenerating axons were surrounded by common Schwann cells to form small bundles, with some axons at the periphery being partly in direct contact with alginate. At the distal stump, numerous Schwann cells had migrated into the alginate 8–14 days after surgery. They had no basal laminae. The diameter of regenerated myelinated fibers was small (approximately 1 μm) at 8 weeks, but increased in diameter, having a distribution pattern similar to that of normal nerve 21 months after surgery. Much better nerve regeneration was found in alginate gel-, than collagen sponge-, and fibrin glue-implanted distal stump 12 months after surgery. These results indicate that alginate gel has good biocompatibility for regenerating axon outgrowth and Schwann cell migration, and that regenerated fibers can have a diameter as thick as that of normal fibers in the long term. Alginate gel is a promising material for use as an implant for peripheral nerve regeneration. Electronic Publication     

Contextual effects on smooth-pursuit eye movements

Segregating a moving object from its visual context is particularly relevant for the control of smooth-pursuit eye movements. We examined the interaction between a moving object and a stationary or moving visual context to determine the role of the context motion signal in driving pursuit. Eye movements were recorded from human observers to a medium-contrast Gaussian dot that moved horizontally at constant velocity. A peripheral context consisted of two vertically oriented sinusoidal gratings, one above and one below the stimulus trajectory, that were either stationary or drifted into the same or opposite direction as that of the target at different velocities. We found that a stationary context impaired pursuit acceleration and velocity and prolonged pursuit latency. A drifting context enhanced pursuit performance, irrespective of its motion direction. This effect was modulated by context contrast and orientation. When a context was briefly perturbed to move faster or slower eye velocity changed accordingly, but only when the context was drifting along with the target. Perturbing a context into the direction orthogonal to target motion evoked a deviation of the eye opposite to the perturbation direction. We therefore provide evidence for the use of absolute and relative motion cues, or motion assimilation and motion contrast, for the control of smooth-pursuit eye movements.