REFERENCE FRAMES FOR VISUO MOTOR CONTROL

空间记忆理论强调自我中心的参照 框架作为中央存储库。 相反，我们的工作使用过滤器 在皮层细胞中发现的反应表明， 存储空间目标的方法。 目标按其要素存储 并且这些特征在图像中的位置可以被快速地 只要目光保持不变，无论目光发生变化，都会恢复 可见 这个概念到目前为止只在静态图像上进行了测试 这些都是离线处理的 我们的计划是彻底测试这个 在我们的机器人模拟器上实现它。 这将考验 工作机制的可行性随时间变化的图像， 受到模糊和比例变化影响。 我们最初的重点是 注视控制研究将整合几种不同的注视 方法. 我们的第一个努力将是整合住宿和 聚散度摄像机移动。 目前仅聚散度电路 需要几乎所有可用的视频处理资源。 的 拟议的扩大将允许同时处理几个 不同的视觉功能。

Theories of spatial memory have stressed egocentric references frames as a central repository. In contrast our work using filter responses such as those found in cortical cells suggests an different way of storing spatial targets. Targets are stored by their features and the location of these features in the image can be quickly recovered regardless of changes in gaze as long as they remain visible. This concept has so far only been tested on static images that are processed off-line. Our plan is to definitively test of this scheme by implementing it on our robot simulator. This would test the feasibility of the mechanism working on time-varying imagery that is subject to blur and variations in scale. The initial emphasis of our gaze control research will be to integrate several different fixation methods. Our first effort will be to integrate accommodation and vergence camera movements. Currently the vergence circuit alone requires almost all of our available video processing resources. The proposed expansion would allow concurrent processing of several different visual functions.