Product design is a complicated activity that is highly reliant on individual impressions, feelings and emotions. Back-propagated neural networks have already been applied in Kansei engineering to solve difficult design problems. However, artificial neural networks (ANNs) have a slow rate of convergence, and find it difficult to devise a suitable network structure and find the global optimal solution. This study developed an ANN-based predictive model enhanced with a genetic algorithm (GA) optimization technique to search for close-to-optimal sports shoe color schemes for a given product image. The design factors of the sports shoe were set as the network inputs, and the Kansei objective value was the output of the GA-based ANN model. The results show that a model built with three hidden layers (28 × 38 × 19) could predict the object value reliably. The R2 of the preference objective was equal to 0.834, suggesting that the developed model is a feasible and efficient tool for predicting the objective value of product images. This study also found that the prediction accuracy for shoes with two colors was higher than that for shoes with only one color. In addition, the prediction accuracy for shoes with a relatively familiar shape was also higher. However, the prediction of color preferences is relatively difficult, because the respondents had different individual color preferences. Exploring the sensitivity and importance of the visual factors (form, color, texture) for various image words is a worthy topic for future research in this field.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.