A study on unified organization of visual features in object recognition and spatial vision

物体识别和空间视觉中视觉特征的统一组织研究

基本信息

批准号：
08301007
负责人：
EJIMA Yoshimichi
金额：
$ 5.5万
依托单位：
KYOTO UNIVERSITY
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (A)
财政年份：
1996
资助国家：
日本
起止时间：
1996 至 1998
项目状态：
已结题

项目摘要

The present study focused the binding problems of visual perceptual spatio-temporal features into object recognition and spatial vision. Using psychophysical experiments and brain imaging method, we investigated the following.1) Temporal integration in generation of contour perceptionUsing the sequential presentation of two pairs of disks with a sector removed, we investigated the temporal integration properties of subjective contour perception. It was found that the temporal integration time for subjective contour was longer than that for real contour (edge or line contour), suggesting that the retention of local information takes an important role in generation of subjective contour.2) Binocular integrationWe presented a new illusory perception of motion tranceparency caused by the binocular presentation of counter-phase flickering stimuli. It was suggested that the visual system decomposes the moving images into Fourier components, and that the decomposition may be mediated by the first order motion system such as the direction- and disparity-sensitive motion energy detection system.3) Interaction between moving edge and statinary chromatin edgeWe presented a new phenomina involed perceptually unified organization of a atationary chromatic pattern and a moving black outline or dot pattern. It was suggested that the visual system effectively uses information from moving features to define the shape of overlapping chromatic image regions.4) The others concearned 'The 2-D motion assimilation', 'Summation between nearby motion signals and facilitatory/inhibitory interaction between distant motion signals', 'Interaction between luminance and color cues in generation of neon color spreading', 'A new method for magnetoencephalography', and 'Magnetic responses of human cortical area V1 to sinusoidal grating stimuli.

本研究将视觉感知时空特征的结合问题集中在对象识别和空间视觉上。使用心理物理实验和脑成像方法，我们研究了以下内容。1）在消除了一个扇区的两对磁盘的顺序呈现中，消除了两对磁盘的时间整合，我们研究了主观轮廓感知的时间整合性能。发现主观轮廓的时间整合时间比真实轮廓（边缘或线轮廓）的时间更长，这表明保留本地信息在产生主观轮廓中发挥了重要作用。2）双眼整合我们对动态tr tranceparention的新型虚幻感知引起，这是由反复刺激性刺激性刺激性刺激的双眼呈现引起的。有人建议，视觉系统将移动图像分解为傅立叶组件，并且可以通过一阶运动系统（例如方向和差异敏感的运动能量检测系统）介导的分解。3）3）移动边缘和统计染色质之间的相互作用，涉及一个新的势态在感知统一的动态镀铬模式和动态式的外套或DOT。 It was suggested that the visual system effectively uses information from moving features to define the shape of overlapping chromatic image regions.4) The others concearned 'The 2-D motion assimilation', 'Summation between nearby motion signals and facilitatory/inhibitory interaction between distant motion signals', 'Interaction between luminance and color cues in generation of neon color spreading', 'A new method for magnetoencephalography', and 'Magnetic responses of human皮质区域V1到正弦光栅刺激。