The formation of visual objects

视觉对象的形成

• 批准号: 7037390
• 负责人: JACOB FELDMAN
• 金额: $ 21.56万
• 依托单位: RUTGERS, THE STATE UNIV OF N.J.
• 依托单位国家: 美国
• 财政年份: 2005
• 资助国家: 美国
• 起止时间: 2005-04-01 至 2008-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7037390
• 关键词: clinical research; computational neuroscience; cues; form /pattern perception; human subject; mathematical model; mental process; motion perception; neural information processing; neuropsychological tests; neuropsychology; psychophysics; space perception; statistics /biometry; time perception; vision tests; visual depth perception; visual stimulus; visual tracking

DESCRIPTION (provided by applicant): Perceptual grouping is the process by which the initially raw and inchoate visual image is organized into perceptual "objects". What spatial factors induce perceptual grouping? What is the sequence of computations whereby the image is progressively organized? One source of difficulty in modeling this process is that, unlike many aspects of early vision, perceptual grouping inherently involves non-local: computations - integration of cues from potentially distant locations in the image. Another difficulty in understanding perceptual grouping has been the lack of objective and temporally precise methods for actually measuring the observer's subjective organization of an image. This proposal seeks to combine (a) recent advances in understanding the non-local computations involved in perceptual grouping with (b) novel experimental methods for determining subjective organization. The experimental methods are based on the finding that perceptual objects enjoy certain objectively measurable benefits, including more efficient visual comparisons within them than between distinct objects. This proposal seeks to use this effect to discover what the visual system in fact treats as a perceptual object, and how this percept develops over the course of processing. Most of the proposed experiments involve carefully constructed artificial stimuli with various grouping cues in force, designed to allow detailed comparisons of the strength, interaction, and time-course of each potential grouping cue. In addition, several experiments involve natural images, in order to uncover how perceptual organization proceeds under more naturalistic conditions. This research may lead to technological advancement in the area of computer vision, as well as to better understanding of disorders of perceptual organization such as visual agnosia and dyslexia.

描述（由申请人提供）：感知分组是将最初原始和不成熟的视觉图像组织成感知“对象”的过程。什么样的空间因素导致知觉分组？图像被逐步组织的计算顺序是什么？建模这个过程的困难之一是，与早期视觉的许多方面不同，感知分组本质上涉及非本地计算-整合图像中潜在遥远位置的线索。理解感知分组的另一个困难是缺乏客观和时间上精确的方法来实际测量观察者对图像的主观组织。该建议试图将联合收割机（a）在理解感知分组中涉及的非局部计算方面的最新进展与（B）用于确定主观组织的新颖实验方法相结合。实验方法是基于这样的发现，即感知对象享有某些客观可测量的好处，包括它们内部比不同对象之间更有效的视觉比较。这个提议试图利用这种效应来发现视觉系统实际上把什么当作知觉对象，以及这种知觉在处理过程中是如何发展的。大多数拟议的实验都涉及精心构建的人工刺激，其中包含各种有效的分组提示，旨在详细比较每个潜在分组提示的强度、相互作用和时间过程。此外，一些实验涉及自然图像，以揭示知觉组织如何在更自然的条件下进行。这项研究可能会导致计算机视觉领域的技术进步，以及更好地理解知觉组织障碍，如视觉失认症和阅读障碍。