Shiffrar M, Lorenceau J
Center for Neuroscience, Rutgers University, Newark, NJ 07102, USA.
Vision Res. 1996 Jul;36(14):2061-7. doi: 10.1016/0042-6989(95)00283-9.
Accurate interpretations of image require the segmentation of motion signals produced by different objects with the simultaneous integration of motion signals produced by the same object. We investigated a motion integration paradigm in which the direction of an object's motion could only be determined from an integration of motion signals across the disconnected object edges. In a series of experiments we show that observers' ability to determine object motion depends significantly upon stimulus duration, luminance contrast and edge width. These effects suggest that the visual system, after some delay, relies upon relatively thick, luminance defined contour discontinuities to segment moving images.
对图像的准确解读需要对不同物体产生的运动信号进行分割，同时整合同一物体产生的运动信号。我们研究了一种运动整合范式，在这种范式中，物体运动的方向只能通过跨不连续物体边缘的运动信号整合来确定。在一系列实验中，我们表明观察者确定物体运动的能力显著取决于刺激持续时间、亮度对比度和边缘宽度。这些效应表明，视觉系统在经过一段时间延迟后，依赖相对较粗的、由亮度定义的轮廓不连续性来分割运动图像。
