Development of an Intelligent Model to Predict Tactile Properties from Visual Features of Textile Products

The development of virtual simulation techniques and digitized interactive platforms makes the purchasing of textile products as simple as a click. But some of our real desires about a fabric is still difficult to be satisfied, such as touch. So far, much research has been dedicated to providing consumers with a realistic simulation of fabric through virtual experiences. In the previous study, we have proved that fabrics' tactile properties could be well perceived through products' visual representations. On this basis, the present study is aimed to further study the mathematical model of predicting fabric tactile properties from samples' visual features. Two sensory experiments have been designed to extract the tactile properties and visual features of a set of textile samples represented in video clips. An intelligent method based on rough set and fuzzy set theories has been developed to extract principal visual features for the interpretation of each tactile property. Then a fuzzy neural network has been designed to build a mathematical model between each tactile property and the corresponding principal visual features. Extra experiments have been carried out to verify the effectiveness of the proposed model. The results have proved that it is possible to predict fabric tactile properties from samples' visual features with satisfactory accuracy.