Research On Tactile Feature Modeling And Extraction Of 2D Image

Electrostatic tactile display technology produces tactile perception of the shape and texture, achieving visual and force tactile feedback synchronously. It becomes a research hotspot in tactile interaction and widely applies in complementary medicine, education, virtual reality etc. The modeling and extraction of image tactile features is one of key points in electrotstatic tactile system to achieve tactile feedback. However, existing image tactile rendering algorithms are proposed to feedback relative space position of shape aiming at mechanical haptic apparatus. With the development of electrostatic tactile technology and digital medium, a real-time and realistic tactile rendering model on touchscreen may become the inevitable requirement in the field of tactile interaction. Research on realistic, low complexity, generic rendering model of image tactile feature, not only enhances immersion and interactivity for electrostatic tactile display but also owns important application value.This paper focuses on modeling and extracting of 2D image tactile feature for electrostatic tactile display system. As two-dimensional images lacks structural information to simulate force tactile feedback. Firstly, this paper researches on extracting and enhancing algorithms of image shape and texture. It extractes height, gradient, enhanced texture and local texture feature through utilizing the method of shape from shading, fractional differential, frequency domain decomposition and local Fourier transform. Then, based on the principle of electrostatic tactile generation, a rendering model between image feature and actuated signal has been proposed to create tactile force. An electrostatic tactile rendering model for convex shape has been proposed to reproduce shape tactile. To further enhance the sensations of electrostatic tactile, an electrostatic tactile rendering model based on enhanced texture detail is established, realizing more exquisite realistic texture tactile feedback. At last, validate the availability of electrostatic tactile rendering models through designing perceptive experiments.The main contributions and innovative work of this paper are as follows:(1) Analyse the principle and system components of the electrostatic tactile display aiming at simulating precise electrostatic tactile. Research on extracting and enhancing algorithms of image shape and texture, including shape from shading, frequency domain decomposition, texture enhancement based on adaptive fractional differential method, texture feature extraction based on local Fourier transform, receiving accurate 3D shape and clearer texture details. Analyse the precision and accuracy of shape and texture by simulation which provide a theoretical basis for electrostatic tactile systems. Meanwhile, define roughness and contrast according to three-dimensional shape to avaluate image, provide a reference to establish typical electrostatic tactile library of image.(2) Design and implement electrostatic tactile rendering model based on shape height, gradient, texture details faeture. These rendering models can extract accurate shape, contour, texture height, establish proportial mapping relationship between image features and electrostatic force. Combine the psychophysical model between electrostatic force and actuated voltage of different amplitudes and frequencies, this paper builds a mapping process between image feature and tactile actuated signal. Therefore, construct a whole process of image electrostatic tactile rendering.(3) For a typical convex shape, shape from shading method is utilized to represent exact height and shape. Analyse the forces of finger sliding over real objects and virtual objects, an electrostatic rendering model for shape height and gradient has been proposed to represent convex sensations. In order to enhance texture tactile feedback, establish a high definition texture tactile rendering based on texture enhancement. It utilizes fractional order differentiation to adaptively enhance texture feature according to image gradient.. In order to feedback the finer texture, local Fourier transform is used to characterize the local texture by local Fourier coefficient. Then, an electrostatic tactile rendering method based on local texture feature has been established.(4) To verify the effectiveness of the four tactile rendering methods, construct interaction interface of perceptive experiment. Design the process for perceptual experiment and compare force characteristics of four tactile rendering methods. Quantitatively and qualitatively analyse the performance of tactile feedback and the applicability of different image tactile model, according to tactile rendering algorithms, image features and tactile force features.The electrostatic tactile rendering methods of image feature that this paper proposed provide clearer texture and more precise sensation of shape. It can provide theoretical and practical reference for the tactile feedback in electrostatic tactile display technology.