Describing Clothing by Semantic Attributes: A Heuristic Approach

Recognition of clothing categories is an appealing study to emerging applications such as computer aided fashion design for mass customization and e-commerce. In the study, the description and classification of clothing styles is established based on the preference index heuristic. Specifically, we take advantage of consumer’s visual perception of clothing and their free choice description to; understand, characterise and, classify clothing styles. Evaluated on a dataset with 60 style samples of male clothing, the procedure demonstrates promising results in recognizing five clothing categories. Keyword: Clothing, Semantic Attributes, Heuristics, Styles, Computer-Aided Fashion Design