基于突触连接视通路方位敏感的图像分级边缘检测方法

视觉通路上的多级方位敏感特性对于视觉轮廓感知起着关键作用,将为更高层次的视皮层图像理解提供重要的特征信息。从视觉方位敏感机制出发,提出一种图像边缘检测的新方法。利用神经节细胞以及外膝体神经元感受野向心分布的生理结构特性,构建具有突触连接和多方向敏感特性的视皮层下功能层,融合多方向上的神经元脉冲发放信息,将视觉激励映射为边缘敏感图像;构建具有去最优方位感受野特性的初级视皮层的功能层,对前级结构生成的脉冲序列按时间信息进行神经编码,经过感受野内侧向抑制和阈值处理,获得边缘检测结果。对层次模糊而细节丰富的菌落图像进行处理,并以边缘置信度和重构相似度以及两者的加权和作为边缘检测评价指标。结果表明,该方法在完整检测图像边缘的同时,并不引入纹理噪声,有着明显的优势,其对12幅图像的加权和指标均值为0.746 8,显著高于其他对比方法。所提出的方法可以模拟视通路中初级视皮层及视皮层下的方向敏感特性,提供一种基于视觉机制的图像处理和理解新思路。     

基于视觉感光层功能的菌落图像多强度边缘检测研究

借鉴视觉神经系统在轮廓感知中的独特优势,提出一种基于视觉感光层功能的图像边缘检测新方法。构建以带漏感的积累发放（LIF）神经元电生理模型为基本单元的神经元网络;根据特定时间窗口内各个神经元的脉冲发放情况,对神经元的增强(ON)或抑制(OFF)类别进行判断;通过拮抗式感受野特性以及神经元激励的反馈增强模式,实现弱边缘的凸显;为克服视觉感光层所具有的适应性并凸显弱细节的对比度,对图像进行多方向、多距离尺度的移动,并融合感光层神经元网络脉冲发放率的差异信息,最后实现图像边缘的有效检测。以具有丰富边缘特性的20幅菌落图像为样本,以边缘置信度和重构相似度作为评价指标,对多强度边缘进行检测。结果表明,所提出方法可以有效完整地检测出图像多强度边缘,且其对弱边缘检测的重构相似度均值高于08,检测准确性有显著的提高(P<005)。所提出的利用生理视觉系统特性进行边缘检测,为包含多强度边缘信息的图像处理提供崭新的思路。     

基于视通路多感受野朝向性关联的轮廓检测方法

基于视通路中各层次的感受野特性以及层次间感受野的关联特性,提出一种图像轮廓检测新方法。根据神经元的放射状树突接收特性,构建检测空间差异性信息的视网膜神经元网络,对轮廓信息进行前级编码;建立LGN细胞的非经典感受野调节机制,结合感受野的层次变换特性,对前级编码结果进行全局调节;提出简单细胞的多感受野朝向性关联模型,模拟初级视皮层简单细胞的方向选择特性;融合多个方向上的轮廓响应,经过非极大值抑制和阈值处理,得到轮廓检测结果。以RuG图库40幅图片为实验对象,检测结果与基准轮廓图的平均P指标为0.43,结果显示该方法能够有效凸显主体轮廓,强化轮廓与纹理边缘的区分度。利用多感受野的层次关联特性,实现图像轮廓信息的编码与检测应用,为研究高级视皮层的图像理解和视觉认知提供新的思路。     

基于抑制性突触多层神经元群放电编码的图像边缘检测

图像边缘检测所获得的目标轮廓细节质量,对于后续图像分析或理解过程具有重要作用。提出一种基于视觉机制的图像边缘检测新方法:构建抑制性突触连接的多层神经元群,在待检测图像的激励下,分析7像素×7像素视觉感受野窗口内互连接神经元的脉冲放电过程,记录发放时刻进行次序编码;同时考虑神经元之间的侧向抑制作用,在选择注意机制作用下获得增强图像;之后利用Log Gabor滤波器模拟视觉系统中的方向选择特性,获取8个方向的滤波结果,经过输出层神经元群的融合处理并经灰度映射到0～255的范围后获得边缘图像。对含有丰富边缘细节特性的24幅菌落图像进行处理,其处理结果的ROC评价指标均值为0.698 4,优于PCNN法的0.659 3;从评价指标的均方差来看,该结果具有更好的一致性。另外,从信息熵评价指标来看,该方法同样具有一定的优势,能够有效提取图像的边缘信息,而且也能反映更多层次的图像细节。所提出的方法为利用视觉生理特性进行图像处理提供了崭新而有效的思路。     

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。     

用菌紫质LB膜模拟视网膜神经节感受野的零交叉滤波特性

应用菌紫质ＬＢ膜构建了一维ＯＦＦ－中心型视网膜神经节细胞感受野，测定了此人工感受野的周边区和中心区对时域方波光信号的响应。结果表明，此人工感受野的周边区和中心区分别具有动物感受野的ＯＮ－微分响应和ＯＦＦ－微分响应特性。在此基础上，又用此人工感受野模拟了动物神经节细胞感受野的零交叉特性，模拟结果与生物实验结果符合。利用菌紫质ＬＢ膜本身的特性构建的这种人工视网膜神经节感受野，避免了电子学和光学模拟的复     

基于神经元颜色拮抗与动态编码的轮廓检测方法

基于神经元的颜色拮抗特性及神经元群体的动态编码机制,实现对图像的轮廓检测。模拟视皮层下神经元的颜色单拮抗特性,引入单拮抗感受野的动态调节机制,以充分响应颜色边界和亮度边界;利用单细胞的树突极性分布,构建初级视皮层的双拮抗神经元网络,实现对特定方位的视觉刺激响应,有效提取目标轮廓;在神经元的群体感受野内,考虑神经元的动态突触连接,融合单细胞的脉冲频率响应,实现对纹理信息的抑制作用。以BSDS500图库的图像为实验对象,结果显示该方法在提取主体轮廓的过程中能有效抑制纹理信息,其对100幅图像最佳检测结果的P值指标均值和标准差为0.58±0.04,相对CORF和CO等其他对比方法,可提高轮廓提取的准确率。所提出方法可有效实现图像的轮廓检测,为利用颜色信息以及神经元之间的动态编码、实现更高级皮层的图像理解或者视觉认知提供新的思路。     

模拟神经反馈机制的显著性计算模型

根据视觉皮层神经元的非经典感受野的调制特性,构造了模拟视觉神经网络反馈机制的显著性计算模型。该模型充分利用神经网络中的反馈机制,模拟视觉皮层神经元之间侧连接所引起的易化和抑制作用,对视频图像序列中全局运动造成的背景噪声进行抑制,从而获得视频运动对象的显著图,提取完整的运动目标。实验结果表明,该计算模型能较好地抑制视频图像序列的全局运动,获得运动对象的显著性图,有效地实现移动镜头下的运动目标检测,且显著性的计算性能高于其它计算方法。     

医学图像的显示修正

医学图像的显示效果对临床诊断非常重要.本文根据人眼的视觉特性,提供了一个对医学图像显示进行修正的方法.通过应用本方法不仅可以提高图像的显示效果,而且可以使同一幅图像在不同的系统上得到视觉上一致的显示.     

一种模拟视觉机制的图像分割模型

提出一个小波域多尺度马尔柯夫随机场模型用于模拟视觉系统在图像分割中的若干功能。针对人类视觉系统具有特征检测器、等级层次性、双向连接性、学习机制等功能,对输入场景,该模型用小波变换提供该场景图像的稀疏表示,模拟特征检测器功能;用金字塔结构模拟等级层次性;用两类信息流模拟双向连接性,分别刻画自底向上的输入图像特征提取过程以及自顶向下的反馈过程;用迭代过程模拟学习机制;采用多尺度马尔柯夫随机场模型实现图像分割。实验表明,该模型对真实采集到的不同类型的生物医学图像进行分割,取得优于一些传统分割算法的结果。     

GABA(B) receptor-mediated modulation of cutaneous input at the cuneate nucleus in anesthetized cats

This study examined the modulatory influence exerted by GABA(B) receptors on the transmission of cutaneous afferent input to cuneate nucleus neurons in anesthetized cats. Electrical stimulation at the center of a receptive field activated cuneate nucleus cells at latencies of < or = 7 ms whereas stimulation at neighboring sites (receptive field edge) increased the response latency. Extracellular recording combined with microiontophoresis demonstrated that GABA(B) receptors are tonically active. Blockade of GABA(B) receptors prolonged sensory-evoked response durations and decreased times of occurrence of successive bursts whereas the agonist baclofen suppressed both these effects. Ejection of baclofen delayed the evoked response from the receptive field edge with respect to the receptive field center response and inhibited responses from the receptive field edge more effectively than responses from the receptive field center. From these results it is concluded that activation of GABA(B) receptors precludes cuneate cells from reaching firing threshold when afferent inputs are weak, spatially modulate cuneate nucleus excitability, play a major role in temporal pattern of discharges, and shape cutaneous receptive fields.     

Responses of single units in the monkey superior colliculus to moving stimuli

1. Single unit responses of pan-directional cells to moving and stationary flashing stimuli were studied in the superficial layers of the superior colliculus in paralysed, anaesthetized rhesus monkeys. The aim of this study was to see how far cell responses to moving stimuli fit in with what would be expected from their responses to stationary flashing stimuli. 2. Both the leading and the trailing edge of a moving stimulus evoke a transient response. If the diameter of moving light spots is increased the strength of the leading edge response increases, reaches a maximum and decreases to a constant value which is similar to the behaviour of the on response when the diameter of flashing spots is increased. The strength of the trailing edge response increases and reaches the same strength as that of the leading edge response. If the width of a long moving slit is increased, the strength of the leading edge response is the same at all slit widths, while the strength of the trailing edge response shows a course similar to that of the trailing edge response if the spot diameter is increased. If the length of a wide moving slit is increased both the leading and the trailing edge responses decrease. These results indicate that the strength of both leading and trailing edge responses is dependent on the degree the inhibitory surround is activated. 3. The leading and the trailing edge of a stimulus evoke their responses at the same position in the receptive field independent of the direction of movement. 4. Increasing the velocity of a moving stimulus shows that in general the leading edge response is present up to higher velocities than the trailing edge response independent of the sign of contrast. The burst duration to moving stimuli decreases with increasing stimulus velocity and appears to be determined by the time a moving edge is present in the receptive field centre. When this time becomes shorter than 10--20 ms, the burst duration for moving stimuli is constant and about the same as for flashing stimuli. This indicates that, although spatial receptive field properties can vary considerably, temporal receptive field properties show a strong similarity among different units. 5. The response latencies to light and dark moving edges are the same, which in turn are about equal to the response latencies to stationary flashing stimuli. 6. Stimulation experiments show that the general response characteristics to moving stimuli can be predicted by using a set of receptive field parameters derived from responses to stationary flashing stimuli. The most important variable of moving stimuli appears to be the period of time a moving contour is present within the receptive field centre, besides the degree of activation of the inhibitory surround.     

The speed of context integration in the visual cortex

The observation of figure-ground selectivity in neurons of the visual cortex shows that these neurons can be influenced by the image context far beyond the classical receptive field. To clarify the nature of the context integration mechanism, we studied the latencies of neural edge signals, comparing the emergence of context-dependent definition of border ownership with the onset of local edge definition (contrast polarity; stereoscopic depth order). Single-neuron activity was recorded in areas V1 and V2 of Macaca mulatta under behaviorally induced fixation. Whereas local edge definition emerged immediately (<13 ms) after the edge onset response, the context-dependent signal was delayed by about 30 ms. To see if the context influence was mediated by horizontal fibers within cortex, we measured the latencies of border ownership signals for two conditions in which the relevant context information was located at different distances from the receptive field and compared the latency difference with the difference predicted from horizontal signal propagation. The prediction was based on the increase in cortical distance, computed from the mapping of the test stimuli in the cortex, and the known conduction velocities of horizontal fibers. The measured latencies increased with cortical distance, but much less than predicted by the horizontal propagation hypothesis. Probability calculations showed that an explanation of the context influence by horizontal signal propagation alone is highly unlikely, whereas mechanisms involving back projections from other extrastriate areas are plausible.     

基于空洞空间金字塔池化的U-Net网络在肺部图像分割上的应用

目的：胸部X线图像中肺野的自动分割是相关疾病筛查和诊断的关键步骤,为了适应计算机辅助诊断系统的要求,提出一种基于空洞空间金字塔池化的U-Net网络对胸部X线图像中肺野进行自动分割。方法：在编码和解码之间引入带有空洞卷积的空间金字塔池化用于扩大接受域;同时,在多个尺度上获取图像上下文信息,用于从胸片中分割肺野,使用Montgomery数据集及深圳数据集进行验证。根据医学图像分割常用指标准确性、Dice相似系数及交并比评价基于空洞空间金字塔池化的U-Net网络分割肺野的性能。结果：验证准确性为98.29%,Dice相似系数为96.61%,交并比为93.47%。结论：本文提出一种基于空洞空间金字塔池化的U-Net网络用于分割肺野,相较于其他方法学习到更多边缘分割特征,取得更好的分割结果。     

Receptive fields of simple cells in the cat striate cortex

1. The excitatory and inhibitory components in the receptive fields of unimodal simple cells in the striate cortex of the cat anaesthetized with nitrous oxide have been described using slits of light and single light-dark edges as stimuli.2. There is a small excitatory region (excitatory complex) centrally located in the receptive field that is made up of various combinations and spatial arrangements of subliminal excitatory and discharge subregions or centres.3. The subliminal excitatory centres were revealed by a binocular facilitation technique. The excitability of the cell was raised by repeated stimulation via one eye while the neurone was tested with single edges via the other eye.4. The subliminal excitatory and discharge centres are each specifically activated by only one type of edge, light-dark or dark-light, and then only in one direction of motion. All the subregions in the excitatory complex have the same optimal stimulus orientation.5. Inhibitory components in the receptive field were identified by stimulating the cell with bars of light and single edges against an artificial background discharge produced by repeated stimulation separately applied either to the same eye (monocular conditioning) or to the other eye (binocular conditioning). There are powerful inhibitory sidebands to either side of the excitatory complex and these inhibitory regions merge to include the excitatory complex when stimulus orientation is angled away from the optimal.6. Excitation is highly stimulus specific whereas inhibition is non-specific.7. The organization of the two receptive fields of a binocularly discharged cell can be closely similar.8. The attempt is made to translate the concept of subliminal excitatory and discharge centres into specific neural mechanisms involving both the geniculo-cortical input and various intracortical circuits.9. These new developments call for only minor modifications to the model we have proposed for the organization of the receptive field.     

Some evidence concerning the physiological basis of the periphery effect in the cat's retina

Summary Peripheral regions of retinal ganglion cell receptive fields were investigated in cerveau isolé cats using automatically presented moving stimuli. By presenting appropriate stimulus configurations it was shown that the excitatory response to peripherally moving objects (the periphery effect) is not due to an extension of receptive field surround properties into the periphery. The differential effects of solid and striped stimuli indicated that the effectiveness of an object in eliciting the periphery effect is related to the amount of moving edge which it presents. Arguments are presented that the periphery effect results from interactions between spatially separated receptive field organizations. Certain difficulties which these findings raise with respect to the current definition of receptive field are discussed.