Product Design Aesthetic Database

A total of 10,003 images of product design submissions of Hardware Product Design Competitionwere collected to form the product design image database, which has 835 samples of awarded class (high aesthetic level), 4,990 samples of qualified class (middle aesthetic level) and 4,178 samples of eliminated class (low aesthetic level). The Product Design Aesthetic Database includes dataset of image features extracted by the optimal method VGG-19.

Image features, rather than product design images, were provided in the dataset due to design copyright protection. The labels of the classes of awarded/ qualified/ eliminated were listed as the last row in the csv dataset document. In the dataset csv file, samples of awarded class are labeled as “nominees”, and samples of qualified class are labeled as “middle class”, andsamples of eliminated class are labeled as “eliminated”.