触觉二维图像识别中2D-3D空间信息转换的认知机制

可触摸的触觉二维图像是视觉障碍人群获取图像信息的重要方式。目前大多数触觉二维图像都是直接由视觉二维图像转化为的可触摸线条图。在视觉二维图像中,通常运用透视和视角等视觉原理将三维空间关系转换为二维平面关系。视觉系统经过长期大量知觉学习,习得了这种二维到三维的映射关系。但是触觉识别二维图像时,触觉系统如何建立二维平面与三维空间的映射,目前尚有待进一步的研究。影响触觉识别二维图像中二维-三维空间信息转换的视觉因素主要有透视、视角、遮挡、纹理梯度和镂空,直接将视觉二维图像转化为的触觉二维图像时,图像中包含的上述视觉因素通常会干扰触觉识别。结合已有研究,试图提出“双表象加工模型”来解释触摸二维图像时二维到三维空间信息转换的认知机制。该模型认为触觉识别二维图像依赖于两个表象系统的整合,即物体表象系统(涉及物体的大小、形状和纹理)与空间表象系统(涉及物体的空间关系、透视和视角)。两种表象系统的信息最终进行整合,在物体表象和空间表象成功匹配的基础上建立二维图像与三维空间之间的映射,通达长时记忆中的三维物体表征。双表象加工模型将有助于我们深入认识触知觉的认知机制,也将为触觉二维图像的设计提供理论依据。

Cognitive mechanisms of 2D-to-3D spatial information transformation in haptic recognition of 2D images

Two-dimensional (2D) tangible image is an important alternative way for visually impaired people to obtain image information. Most existing tactile 2D images are tangible line drawings directly transformed from visual 2D images. The expression of spatial information in the visual 2D images usually follows principles of perspective and viewpoint transforming three-dimensional (3D) space into 2D planar. The mapping from 2D image to 3D space in visual system is learned through long-time perceptual learning. However, it still needs further exploration about cognitive mechanisms of 2D-to-3D spatial information transformation established by haptic system in haptic recognition of 2D images. The visual features of 2D images, including perspective, viewpoint, occlusion, texture gradient and hollow-out, affect the 2D-to-3D spatial information transformation in haptic recognition of 2D images mainly. These visual factors usually interfere with the recognition of 2D tactile images which are directly transformed from 2D visual images. Therefore, visual factors, such as perspective and viewpoint, cannot be directly transferred to tactile 2D images. Based on the findings of existing researches, a dual-imagery processing model is proposed to explain the mechanisms of 2D-to-3D spatial information transformation when touching 2D tangible images, and the cognitive mechanism concerning about how visual factors modulate the 2D-to-3D spatial information transformation. The following are key points of this model: