CHS: Large: Collaborative Research: TextureShop: Tools for the Composition and Display of Virtual Texture

Carnegie Mellon University, Pittsburgh PA

Investigators

Abstract

When we interact with the physical world touch is a vitally important sensory channel, but when we interact with the digital world that is not yet the case. Historically, this situation may have been principally due to inadequate tactile displays, but that limitation is quickly disappearing. Increasingly, the principal limitation is the lack of tactile content. The goal of this collaborative research that involves scientists at three institutions is to empower the content creator, by enabling people to perform the same sorts of operations with tactile textures that they routinely perform with photographs. Those operations include "capturing" a texture, building a mathematical representation of it, creating and displaying synthetic versions that feel very much like the original, enhancing it in various ways (e.g., making it rougher or more velvety), and ultimately "composing" novel textures that nonetheless feel realistic and credible. As a tangible step in this direction, an open source, open hardware project begun under prior NSF support will be continued and expanded. That project resulted in the distribution of surface haptic devices to about a dozen different labs, leading to a variety of research studies. In this project, a low-cost surface haptic display and a variety of applications and software tools will be distributed to about 50 early adopters in the research community. Those individuals will be engaged in this research (e.g., by helping to "tag" various textures), and will be empowered to carry out their own research. In addition, workshops will be organized at major human-computer interaction conferences to support the growing surface haptics community. This work is timely and compelling for a number of reasons. One, scientific understanding of the physical and neuronal bases of texture perception has advanced considerably in recent years. For instance, the relationships between vibrations on the skin (produced when a finger slides across a surface), spike timing in afferent neurons, and high-level percepts such as recognition of a specific texture, have recently been elucidated. Two, "surface haptic" technologies for displaying texture to the bare fingertips have also advanced significantly and can now display complex stimuli across the full bandwidth of tactile acuity. Three, the prevalence of touch screen interfaces has created a plethora of applications such as children's e-books, interfaces for the blind, games, and automobile control panels, which would be well-served by high quality tactile content. The merit of this research is that it will provide a principled foundation for both the creation and manipulation of that content. Contributions will include: the development of a "tactile camera" that is able to capture the relevant frictional and vibratory data from which realistic textures can be recreated; a novel mathematical representation of the salient aspects of texture as well as algorithms for synthesizing artificial textures on the basis of that representation; a suite of techniques for enhancing aspects of texture by direct operation on the mathematical representation, interpolation between multiple textures, and interaction with audio cues; and finally a set of tools for composing novel textures including search, texture combination and scale transformation.