SVM-Based Method for Perceptual Image Recognition

Perceptual image of a product plays a significant role in decision making when users choose a product whose basic function is homogeneous nowadays. Designers try to design products that meet the all kinds of demands of users. However, a big gap between designers and users exists owning to the subjectivity of designers' experience. An objective model to recognize perceptual image of products is proposed by extracting low-level texture features of product pictures. To build up ground truth of perceptual image dataset, 210 chair samples are collected and rated by the students who are majored in industrial design using five-point semantic difference method to three pairs of perceptual image words. Several low-level texture features are explored and extracted. We adopt SVM (support vector machine) to classify perceptual image and SVR (support vector regression) to predict perceptual image score. The experiment results show the model is effective in perceptual image recognition. This model can be applied to help designers understand users' perceptual image better and accelerate design progress.