Mechanism underlying the perception of shape and material properties of light transmitting objects

在这个项目中，我们的目标是研究透光物体的材料和形状感知的几个方面。四个子项目中的前两个专注于薄平面透明层的可感知透射率属性。在这项工作中，我们参考了Faul和Ekroll(2002,2011)提出的感知透明度的过滤模型。第一个子项目的目标是改进、扩展和验证我们最近开发的这个模型的参数空间。这个空间的维度在感知上大致是独立的，在感知上是一致的。这些特性使其非常适合于直观地操作、描述和预测透明层的感知特性。在第二个子项目中，我们研究了光源颜色变化时感知层透过率的稳定性。作为对恒定程度的潜在影响，我们考虑了背景的结构、颜色适应、光照变化的可识别性以及受试者的任务。另外两个子项目的重点是对逼真的三维透明对象的感知。在此类对象中清晰可见的镜面反射和背景扭曲提供了可用于推断对象属性的附加信息源。在第三个子项目中，我们调查了这些信息是否可以用来推断物体材料的折射率。从理论上讲，这似乎是可能的，因为镜面反射的强度和背景扭曲的程度都随着折射率的增加而增加。我们之前用静态失真模式进行的实验导致了否定的结果。在这个子项目中，我们将测试是否可以使用对象相对于场景和/或观察者的运动引起的扭曲图案的变化来估计折射率。此外，我们还考虑了镜面反射强度和全反射图案作为折射率的潜在线索。在第四个子项目中，我们研究了透光对象的特殊情况下的3D形状感知。到目前为止，我们所获得的结果表明，在这种情况下，形状感知依赖于不同的提示，并且通常不如不透明对象的情况下的准确。这个子项目的一个目标是量化形状识别的准确性和精确度如何依赖于场景和目标对象的属性。此外，我们将尝试识别视网膜图像中的规则，这些规则被视觉系统用来识别光传输对象的形状。

In the project we aim to investigate several aspects of material and shape perception of light transmitting objects. The first two of four sub-projects focus on the perceived transmittance properties of thin planar transparent layers. In this work we refer to the filter model of perceived transparency proposed by Faul and Ekroll (2002, 2011). The goal of the first sub-project is to improve, extend and validate a parameter space for this model that we have recently developed. The dimensions of this space are approximately perceptually independent and perceptually uniform. These properties makes it well suited to intuitively manipulate, describe and predict the perceived properties of transparent layers. In the second sub-project we study constancy of the perceived layer transmittance under changes in the color of the illuminant. As potential influences on the degree of constancy we consider the structure of the background, color adaptation, the recognizability of an illumination change, and the subjects' task. The focus of the other two sub-projects is on the perception of realistic three-dimensional transparent objects. Specular reflections and background distortions that are clearly visible in such objects provide additional sources of information that may be used to infer the objects' properties. In the third sub-project we investigate, whether this information can be used to infer the refractive index of the objects' material. In theory, this seems possible because both the strength of specular reflections and the degree of background distortion increase with the refractive index. Our previous experiments with static distortion patterns led to negative results. In this sub-project we will test, whether changes in the distortion pattern caused by object motion relative to the scene and/or the observer can be used to estimate the refractive index. In addition we also consider the strength of specular reflection and the pattern of total reflections as potential cues for the refractive index. In a fourth sub-project we study 3D-shape perception in the special case of light-transmitting objects. The results that we obtained so far indicate that in this case shape perception relies on different cues and is often less accurate than in the case of opaque objects. A goal of this sub-project is to quantify how the accuracy and precision of shape recognition depends on the properties of the scene and the target object. Furthermore, we will try to identify regularities in the retinal image that are used by the visual system to recognize the shape of light transmitting objects.