基于物体偏向注意机制的机器人视觉搜寻方法

针对环境冗余信息下目标整体选择注意及机器人应用问题,提出一种目标驱动的基于物体偏向注意机制的机器人视觉搜寻方法.建立任务目标描述模型,基于高斯混合模型进行特征聚类获得先验知识.引入原对象和偏向特征模版,建立一种基于物体的偏向注意模型实现对整个任务目标的搜索与选择注意,并获得任务相关的显著图实现对潜在目标的跟踪与接近控制,将视觉伺服控制算法扩展到机器人认知领域.实验结果证明了该方法适用于目标指向性的机器人应用领域.     

Early ERP effects on the scaling of spatial attention in visual search

This article describes the examination of the spatial‘scaling＇ effect of visual attention with the technique of event-related potential （ERP）. Eighteen participants were involved in a visual search task in which the cue-target paradigm was used. The search array was three concentric circles consisting of randomly selected English letters that were equally distributed in each circle. The behavioral and ERP data were recorded, respectively. The behavioral results showed that the response time increased and the response accuracy decreased with the increase of precue size. The ERPs amplitude of PI and NI components evoked by search array increased with the reduction of precue size. However, the latencies of these ERP components did not show significant differences between conditions. The hierarchical data of both behavioral assessment and ERPs provided evidence for the spatial ＇scaling＇ effect of visual attention. The amplitudes of PI and N 1 components may be used as indices to examine the effect of spatial ＇scaling＇. In different tasks, the display-set size of stimuli and the task complexity may be important factors that affect the attention allocation.     

Early ERP effects on the scaling of spatial attention in visual search

This article describes the examineation of the spatial ‘scaling’ effect of visual attention with the technique of event-related potential (ERP). Eighteen participants were involved in a visual search task in which the cue-target paradigm was used. The search array was three concentric circles consisting of randomly selected English letters that were equally distributed in each circle. The behavioral and ERP data were recorded respectively. The behavioral results showed that the response time increased and the response accuracy decreased with the increase of precue size. The ERPs amplitude of P1 and N1 components evoked by search array increased with the reduction of precue size. However, the latencies of these ERP components did not show significant differences between conditions. The hierarchical data of both behavioral assessment and ERPs provided evidence for the spatial ‘scaling’ effect of visual attention. The amplitudes of P1 and N1 components may be used as indices to examine the effect of spatial ‘scaling’. In different tasks, the display-set size of stimuli and the task complexity may be important factors that affect the attention allocation.     

Illusory shifts in visual direction accompany adaptation of saccadic eye movements.

A central problem in human vision is to explain how the visual world remains stable despite the continual displacements of the retinal image produced by rapid saccadic movements of the eyes. Perceived stability has been attributed to 'efferent-copy' signals, representing the saccadic motor commands, that cancel the effects of saccade-related retinal displacements. Here we show, by means of a perceptual illusion, that traditional cancellation theories cannot explain stability. The perceptual illusion was produced by first inducing adaptive changes in saccadic gain (ratio of saccade size to target eccentricity). Following adaptation, subjects experienced an illusory mislocalization in which widely separated targets flashed before and after saccades appeared to be in the same place. The illusion shows that the perceptual system did not take the adaptive changes into account. Perceptual localization is based on signals representing the size of the initially-intended saccade, not the size of the saccade that is ultimately executed. Signals representing intended saccades initiate a visual comparison process used to maintain perceptual stability across saccades and to generate the oculomotor error signals that ensure saccadic accuracy.     

Feature-based attention influences motion processing gain in macaque visual cortex.

Changes in neural responses based on spatial attention have been demonstrated in many areas of visual cortex, indicating that the neural correlate of attention is an enhanced response to stimuli at an attended location and reduced responses to stimuli elsewhere. Here we demonstrate non-spatial, feature-based attentional modulation of visual motion processing, and show that attention increases the gain of direction-selective neurons in visual cortical area MT without narrowing the direction-tuning curves. These findings place important constraints on the neural mechanisms of attention and we propose to unify the effects of spatial location, direction of motion and other features of the attended stimuli in a 'feature similarity gain model' of attention.     

The time course of visual categorization： Electrophysiological evidence from ERP

Categorization is the process to ascertain the identity of an object by a specific standard[1,2]. Categorization isa fundamental cognitive process; in fact, the raw per- ception would be of little use without the effective clas- sification of original inf…     

Genetic and neuroimaging studies of cholinergic-neurotrophic modulation of visual attention

Molecular genetics can provide complementary information to neuroimaging methods on the neural mechanisms of perception and cognition. We investigated the role of cholinergic and neurotrophic genes in the modulation of visual attention, as reflected in behavioral, ERP, and MRI measures. Three studies examined the influence of single nucleotide polymorphisms (SNPs) in a nicotinic cholinergic gene, CHRNA4, and a neurotrophic gene, APOE. An association was found between a CHRNA4 SNP and reaction time (RT) measures of orienting of visual attention in space. This behavioral finding was complemented by an association between the same SNP and ERP components reflecting early perceptual/attentional processing. In contrast to these relatively specific effects, the APOE e4 SNP had broad effects on cognition, including visual attention and working memory. An interaction was also found between APOE and CHRNA4 for white matter volume and RT measures of attentional efficiency. These studies confirm the role of nicotinic cholinergic innervation for visual attention and show the benefits of a “multi-level phenotyping” approach in the neurogenetics of attention.     

Genetic and neuroimaging studies of cholinergic-neurotrophic modulation of visual attention

Molecular genetics can provide complementary information to neuroimaging methods on the neural mechanisms of perception and cognition. We investigated the role of cholinergic and neurotrophic genes in the modulation of visual attention, as reflected in behavioral, ERP, and MRI measures. Three studies examined the influence of single nucleotide polymorphisms (SNPs) in a nicotinic cholinergic gene, CHRNA4, and a neurotrophic gene, APOE. An association was found between a CHRNA4 SNP and reaction time (RT) measures of orienting of visual attention in space. This behavioral finding was complemented by an association between the same SNP and ERP components reflecting early perceptual/attentional processing. In contrast to these relatively specific effects, the APOE e4 SNP had broad effects on cognition, including visual attention and working memory. An interaction was also found between APOE and CHRNA4 for white matter volume and RT measures of attentional efficiency. These studies confirm the role of nicotinic cholinergic innervation for visual attention and show the benefits of a “multi-level phenotyping” approach in the neurogenetics of attention.     

基于视觉注意机制的小目标预检测特性

针对一种使用了离散非线性采样的视觉注意机制模型,对其应用在小目标预检测上的检测概率进行了研究和分析．在一定的假设条件下,从理论上推导出了模型中影响小目标预检测概率的各个因素,包括目标大小、目标所在采样节点的大小、输入图像信噪比和目标均匀性参数等．由于采用了非线性的采样模型,注视点的转移将带来不同图像区域空间分辨率的变化,从而不能获得总的检测概率与上述各参数的有意义的关系,通过实验分析了这些因素在小目标预检测中对检测概率的具体影响．     

面阵CCD在视觉测量中的应用

论述面阵CCD传感器的工作原理，借助几个典型示例说明了面阵CCD传感器在视觉测量中的应用原理，并介绍了一套基于面阵CCD的视觉测量系统OVM25．     

Selective deficit of visual search in moving displays after extrastriate damage

A visual cue that is often associated with significant stimuli, such as those provided by prey and predators, is movement relative to the observer. An efficient visual system should be able to direct attention to those parts of the visual field that contain such stimuli. What is needed is a system that can filter by movement difference. This could direct attention to a moving item among stationary items, or an item moving in one direction against a background moving in a different direction. Visual search experiments have shown that people are indeed able to filter by movement; that is, they can attend to just the moving items in arrays of moving and stationary stimuli. Single-cell recordings from monkey visual cortex show that the medial temporal cortical area (MT) has some of the properties required to filter by movement. We have now linked these two observations by showing that a patient with bilateral lesions to the presumed human homologue of MT cannot restrict visual attention to the moving items in arrays of both moving and stationary items. This suggests that MT is the site of a movement filter used in normal visual processing.     

An ERP study for the degrees of the difficulty in visual search

The present study mainly probed the relationships between ERP components and the degrees of difficulty in visual search. Three experimental and one controlled conditions were randomly selected. Results showed that the P1, N1, and P2 amplitudes and latencies had no significant differences among the three experimental and one controlled conditions, and the ERP waveforms of low perceptual load of visual search were different from that of high perceptual load. These results might indicate that the latencies and amplitudes of initial P1, N1 and P2 in high and low perceptual load conditions were almost the same. The ERP waveforms of parallel search were very different from that of serial search, and the ERP waveforms of serial search were very similar. The P3 amplitudes increased as the degrees of the difficulty in visual search decreased.     

An ERP study for the degrees of the difficulty in visual search

The present study mainly probed the relationships between ERP components and the degrees of difficulty in visual search. Three experimental and one controlled conditions were randomly selected. Results showed that the P1, N1, and P2 amplitudes and latencies had no significant differences among the three experimental and one controlled conditions, and the ERP waveforms of low perceptual load of visual search were different from that of high perceptual load. These results might indicate that the latencies and amplitudes of initial P1, N1 and P2 in high and low perceptual load conditions were almost the same. The ERP waveforms of parallel search were very different from that of serial search, and the ERP waveforms of serial search were very similar. The P3 amplitudes increased as the degrees of the difficulty in visual search decreased.     

视觉注意特征的变分水平集图像分割模型

针对传统主动轮廓模型较低的鲁棒性能和对先验知识融合能力的不足,基于视觉注意机制的先验知识和曲线演化的理论框架,首先建立图像底层视觉显著性特征的数学模型,在此基础上提出新的曲线演化能量泛函模型,然后对该能量泛函采用变分水平集方法进行推导,得到曲线演化的偏微分方程,数值实验表明该模型相对于经典主动轮廓模型具有更强的抗噪性与分割效率.该模型的提出为进一步在主动轮廓模型中引入更高层次视觉显著性特征、得到更优越的分割模型打下了基础.     

The role of different cues in the brain mechanism on visual spatial attention

The visual spatial attention mechanism in the brain was studied in 16 young subjects through the visual search paradigm of precue-target by the event-related potential (ERP) technique, with the attentive ranges cued by different scales of Chinese character and region cues. The results showed that the response time for Chinese character cues was much longer than that for region cues especially for small region cues. With the exterior interferences, the target stimuli recognition under region cues was much quicker than that under Chinese character cues. Compared with that under region cues, targets under Chinese character cues could lead to increase of the posterior P1, decrease of the N1 and increase of the P2. It should also be noted that the differences between region cues and Chinese character cues were affected by the interference types. Under exterior interferences, no significant difference was found between region cues and Chinese character cues; however, it was not the case under the interior interferences. Considering the difference between the exterior interferences and the interior interferences, we could conclude that with the increase of difficulty in target recognition there was obvious difference in the consumption of anterior frontal resources by target stimuli under the two kinds of cues.     

线性图标内部特征对视觉搜索的影响机理

图标的视觉搜索在人机交互中起重要作用,线性图标是设计师广泛使用的图标形式。为探究线宽和亮度对比度的变化对线性图标搜索绩效的影响机理,采用经典搜索范式并进行眼动实验验证。实验a结果表明,图标线宽的主效应显著,当图标L_长度/S_宽度比值约为12%时,线性图标的搜索速度最快。实验b1通过具有连续数量级的亮度对比度变化发现：当图标线宽适中时,图标亮度对比度的主效应显著,可通过增加亮度对比度提高图标的视觉搜索效率,亮度对比度是认知绩效的主要影响因素;实验b2采用极小线宽值的线性图标作为亮度对比度实验刺激,结果发现在较高对比度区间内,对比度越高反应时间越长,说明过高的对比度显著影响了极小线宽值图标的视觉搜索,线宽和对比度之间存在交互效应。实验c通过眼动实验从视神经认知角度解释了实验b2中线宽和对比度存在交互效应的结论。研究表明,线性图标的线宽和对比度对视觉搜索效率产生显著影响。     

一种新的基于注意力机制的细粒度视频分析神经网络结构

针对包含细微差异动作的视频数据集,提出了一种用于分辨细粒度差异动作的深度神经网络.该网络结构由一个三维卷积(C3D)网络的轻量化变体和一个基于注意力机制的长短时记忆网络组成,优化了三维卷积网络的深度和注意力机制的权重惩罚项.实验结果表明:该网络可以有效地关注视频中的重要信息,在平均准确率和检测准确率上均有所提升.