An image evaluation approach for parameter-based product form and color design

The parameter-based technique provides an efficient and valid means of constructing 3-D geometric models in many CAD software systems. However, its use is generally restricted to the design of mechanical components with regular configurations, and it is not ideally suited to product form and color design. This paper proposes a rapid conceptual design approach, which creates color-rendered forms and combines parameter-based features with fuzzy neural network theorems and gray theory to predict their image evaluation. Two evaluation models (Evaluation Model I and Evaluation Model II) are developed and applied in a case study of an electronic door lock design. Model I uses a fuzzy neural network to predict the overall image, while Model II uses a gray clustering operation for the color image evaluation and two fuzzy neural networks for the form image evaluation and the overall image evaluation. The results show that the image prediction capability of Model II is superior to that of Model I (RMSE: 0.062 versus 0.105). Furthermore, the overall image evaluation is dominated by the door lock's color rather than by its form (RMSE: 0.071 versus 0.162). The dominance of color in determining the image evaluation may be due to the specified image words, form evolution restrictions, or the membership grade ranges of the test color samples and the test form samples, etc. Having established the superiority of Model II, it is applied to develop a consultative design interface integrated with a professional CAD system in order to demonstrate the effectiveness of the proposed product design and image evaluation approach. The design system presented in this study enables a designer to predict the likely image tendencies of a designed product without the need to create and test a prototype model. Hence, he or she can make any design parameter modifications necessary to ensure that the finished product meets its specified image goals.