Research On Method Of Image-based Product Form Design

With the radical increasing of market competition, product requires more andmore diversities and individualities to meet consumers' demands. While its practicalfunction is no longer the major factor for consumer to consider when purchasing,product's perceptual content is becoming more and more important. Product'sappearance is an important medium to bridge the client and the designer, andproduct's image conveys the implied culture information, which is the effectivemethod to express the client's emotional need. For long time the designer has alwaysrelied on his/her own experience, intuition and inspiration to design model, which isdifficult to satisfy the client's emotional need exactly and efficiently due to thelacking of the effective support from both theory and experiment. Based on the theoryand the method of the Kansei engineering and the artificial intelligence the authorstudied the issue of creative model design orienting to the client's emotional need,realized the structural expressing of the client's emotional need on the image level,performed the associated experiments and analysis, and established the basis of thetheory and technique for the computer-aided image-oriented product model designinnovation. The major research works included in the dissertation are as following:1.Propose a new design idea based on the image of visualization, auditory, andsensitivity: from the angle of visualization, auditory, and sensitivity, the authoranalyzes how to image the model's perception and the method to obtain client'simage, and thus construct a descriptive model to transfer the client's perceptual needsto elements of appearance design. Combining the process to build a model anddescription about quanitization, the author provides the computer expression andapplication mode.2.Verify the user-oriented constructive mode with a car's frontal design as anexample: use multidimensional scaling, principle component analysis, taxonomyanalysis, and etc to extract the product design features and representative semanticsthat affect custom psychological feeling, use multiple regression to formulate therelationship between product design feature and representative semantics. 3.Based on feature orientation, take advantage of genetic neural network andgenetic algorithms to build a product design model: use the concept of featureorientation to make a prototype, and generate a new model with evolution ofinterpolating character parameters. Hence, coordinate with semantic differentialmethod, quantize the user's image on new model, and thus propose an intelligentmethod that combines back-propagation neural network (BPN) and geneticalgorithms to optimize the product design plans according to customer's image.4.Suggest the application of the gray theory on the correlation of product colorand image: build a model on the correlation of gray scales between color combinationvalues, and use a gray clustering method to study the evaluation algorithm on theimage values of color combination plans, and provide a new method for computeraided color design.5.Provide a design mode for product color combination based on the geneticneural network: divide the target product into various color components, and design apoll system using color combination theory, and simulate the consumer's imageevaluation according to the theory of neural net work, and then deliver nextgeneration using genetic algorithm, thus, establish a frame for the system of computeraided product color combination.