Disparity and Correspondence in Human Stereo Vision

人类立体视觉的视差与对应

• 批准号: 1257096
• 负责人: Bart Farell
• 金额: $ 42.76万
• 依托单位: Syracuse University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2013
• 资助国家: 美国
• 起止时间: 2013-03-01 至 2017-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1257096&HistoricalAwards=false
• 关键词: Disparity; Correspondence; Human; Stereo; Vision

Due to their different locations, our left and right eyes have slightly different views of the world. It has been known since the 1830's that a horizontal disparity (shift) between the position of an image on the retinas of the left and right eyes is sufficient for humans to see that image in depth. Combining the two retinal images allows us to see stereoscopically the full three-dimensional volume of scenes (and even some movies and comic books). Stereoscopic vision helps people to navigate, to locate, identify and grasp objects, to judge distances, and to drive vehicles. The goal of this research project is to understand how the brain interprets retinal disparities as cues to stereoscopic depth. Horizontal disparities are relative easy to interpret; this is because of the separation of the eyes, when we are in upright posture, is horizontal. However, in naturalistic scenes disparities can have any direction, not just horizontal, and interpreting them is complicated by the variety of spatial arrangements among objects can influence the disparity directions we encounter. Therefore, understanding how humans see depth requires that we understand how the brain analyzes this two-dimensional (horizontal and vertical) disparity signal. The investigator will measure the depth that people perceive as they view displays containing several patterns, each with its own disparity direction. The knowledge gained from these studies will help scientists understand how humans and other animals combine images from the two eyes in order to recover information not available from either eye alone and how perception of 3-D space can be distorted when this information is combined incorrectly. Learning about the brain's strategies in achieving normal 3-D vision will help explain why visual areas of the brain are organized the way they are and aid in the design of treatments for impaired stereoscopic vision (from, for example, amblyopia and strabismus, or "lazy eye"). It will also help in the design of more effective artificial depth-sensing systems and improve machine-vision algorithms for object recognition, robotics, and visual prosthetic devices.

由于位置不同，我们的左眼和右眼对世界的看法略有不同。自19世纪30年代以来已经知道，图像在左眼和右眼的视网膜上的位置之间的水平视差（偏移）足以使人类深入地看到该图像。结合这两个视网膜图像使我们能够立体地看到场景（甚至一些电影和漫画书）的完整三维体积。立体视觉帮助人们导航，定位，识别和抓住物体，判断距离和驾驶车辆。这项研究项目的目标是了解大脑如何将视网膜差异解释为立体深度的线索。水平视差相对容易解释;这是因为当我们处于直立姿势时，眼睛是水平的。然而，在自然主义场景中，视差可以有任何方向，而不仅仅是水平的，并且解释它们是复杂的，因为物体之间的各种空间排列会影响我们遇到的视差方向。因此，要理解人类如何看到深度，就需要我们理解大脑如何分析这种二维（水平和垂直）视差信号。研究人员将测量人们在观看包含几种模式的显示器时所感知的深度，每种模式都有自己的视差方向。从这些研究中获得的知识将帮助科学家了解人类和其他动物如何将两只眼睛的图像联合收割机组合起来，以恢复单独从任何一只眼睛无法获得的信息，以及当这些信息组合不正确时，三维空间的感知如何扭曲。了解大脑实现正常三维视觉的策略将有助于解释为什么大脑的视觉区域是以它们的方式组织的，并有助于设计立体视觉受损的治疗方法（例如，弱视和斜视，或“弱视”）。它还将有助于设计更有效的人工深度传感系统，并改进用于物体识别、机器人和视觉假体设备的机器视觉算法。