Perception of Object Motion from Optical, Binocular and Scene Information

从光学、双目和场景信息感知物体运动

• 批准号: 7669120
• 负责人: George J Andersen
• 金额: $ 36.35万
• 依托单位: UNIVERSITY OF CALIFORNIA RIVERSIDE
• 依托单位国家: 美国
• 财政年份: 2007
• 资助国家: 美国
• 起止时间: 2007-09-30 至 2012-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7669120
• 关键词: Accidents; Address; Automobile Driving; Cognitive; Complex; Crowding; Data; Detection; Discrimination; Elderly; Environment; Event; Future; Health; Human; Incidence; Knowledge; Length; Metric; Modeling; Motion; Noise; Optics; Perception; Positioning Attribute; Process; Recovery; Reporting; Research; Review Literature; Risk; Safety; Series; Source; Speed; System; Task Performances; Walking; base; falls; object motion; research study; three-dimensional modeling; trafficking

DESCRIPTION (provided by applicant): The purpose of this research is to develop a comprehensive model of the perception of 3D object motion. The recovery of object motion in a 3D environment is important for perceptual tasks such as the detection and avoidance of collisions and performing interceptive actions. The recovery of object motion is also important for cognitive tasks such as planning future events and navigation. Successful performance of these tasks can be important for the health and safety of the observer. For example, successfully detecting and avoiding collisions is an issue in driving safety and accident risk. In addition, a decreased ability to detect and avoid collisions has been shown to be an important factor in the increased risk of falls among the elderly. Understanding the processes involved in the perception of 3D object motion is an important and necessary step in addressing these health issues. We propose that the recovery of 3D object motion is determined by analyses of optical motion, binocular information, and scene information. A series of experiments will be conducted to determine how observers recover the path and speed of an object moving through a 3D scene from these analyses. The types of information to be investigated are: (a) optical motion-the rate of change in projected size and projected position, (b) binocular information, including disparity and vergence, and (c) scene information, including changes in the projected position of an object along a ground plane. In some experiments, noise will be added to one or more sources of information. Observer tasks will include discrimination between different orientations of straight and curved paths, discriminating between straight paths and paths that are bent or curved, and estimating the future position of a moving object. Other experiments will examine the metrics used by an observer in recovering 3D object motion, such as object size, interocular separation, eyeheight, and stride length. The results of these experiments will be used to develop and validate a general model of 3D object motion based on the intrinsic constraints (1C) model. The general model will integrate information from optical motion, binocular, and scene based analyses to estimate the accuracy and bias of human observers in judging the trajectory of objects moving in the 3D world.

描述（由申请人提供）：本研究的目的是开发 3D 物体运动感知的综合模型。 3D 环境中物体运动的恢复对于检测和避免碰撞以及执行拦截动作等感知任务非常重要。物体运动的恢复对于规划未来事件和导航等认知任务也很重要。成功执行这些任务对于观察者的健康和安全非常重要。例如，成功检测和避免碰撞是驾驶安全和事故风险的一个问题。此外，发现和避免碰撞的能力下降已被证明是老年人跌倒风险增加的一个重要因素。了解 3D 物体运动感知所涉及的过程是解决这些健康问题的重要且必要的一步。我们建议通过分析光学运动、双眼信息和场景信息来确定 3D 物体运动的恢复。将进行一系列实验，以确定观察者如何从这些分析中恢复物体在 3D 场景中移动的路径和速度。要研究的信息类型是：（a）光学运动 - 投影尺寸和投影位置的变化率，（b）双眼信息，包括视差和聚散度，以及（c）场景信息，包括投影位置的变化沿地平面的物体。在一些实验中，噪声将被添加到一个或多个信息源中。观察者的任务将包括区分直线和弯曲路径的不同方向，区分直线路径和弯曲或弯曲的路径，以及估计移动物体的未来位置。其他实验将检查观察者在恢复 3D 物体运动时使用的指标，例如物体大小、眼间距、眼高和步长。这些实验的结果将用于开发和验证基于内在约束 (1C) 模型的 3D 物体运动通用模型。通用模型将整合来自光学运动、双目和基于场景的分析的信息，以估计人类观察者在判断 3D 世界中物体移动轨迹时的准确性和偏差。