Abstract
Background: Fabric is one of the keys and vital design factors in fashion design. However, the selection of relevant fabrics is rather complex for designers and managers due to the complexity of criteria at different levels.
Introduction: In this paper, we propose a new fabric recommendation model in order to quickly realize fabric selection from non-technical fashion features only and predict fashion features from any fabric’s technical parameters. This approach is extremely significant for fashion designers who do not completely master fabric technical details. It is also very useful for fabric developers who have no knowledge on fashion markets and fashion consumers.
Methods: The proposed fabric recommendation model has been built by exploiting designers’ professional knowledge and consumers’ preferences. Concretely, we first use fuzzy sets for formalizing and interpreting measured technical parameters and linguistic sensory properties of fabrics and then model the relation between the technical parameters and sensory properties by using rough sets. Next, we model the relation between fashion themes and sensory properties using fuzzy relations. By combining these two models, we establish a hybrid model characterizing the relation between fashion themes and technical parameters.
Results: The proposed model has been validated through a real fabric recommendation case for designer’s specific requirements. We can find that the proposed model is efficient since the averaged value of prediction errors is 8.57%, which does not exceed 10% (generally considered as an allowable range of human perception error).
Conclusion: The proposed model will constitute one important component for establishing an intelligent recommender system for garment design, enabling to support innovations in textile/apparel industry in terms of mass customization and e-shopping.
Keywords: Fashion design, fashion theme, sensory evaluation, fuzzy logic, rough sets, garment design.
[http://dx.doi.org/10.1108/09556221011018658]
[http://dx.doi.org/10.1007/978-3-642-13974-1]
[http://dx.doi.org/10.1016/j.triboint.2012.08.018]
[http://dx.doi.org/10.4028/www.scientific.net/AMR.1048.168]
[http://dx.doi.org/10.21660/2017.39.7350]
[http://dx.doi.org/10.1177/0301006619863264] [PMID: 31337268]
[http://dx.doi.org/10.1007/978-3-7908-1873-4]
[http://dx.doi.org/10.1016/j.ins.2020.05.094]