Article Preview
Top1. Introduction
With the rapid development of computer technology, the combination of computer-aided design technology (Barenghi 2019) and computer-aided production technology have been widely used in the clothing industry and greatly improved the efficiency of clothing design and production (Maldini 2019). However, the computer-aided clothing design and production mainly focus on the provision of a design platform and the storage of design result or the control of a production process (Jhanji 2018). It cannot make a more intuitive definition and evaluation of the clothing design scheme. As a platform for information transmission, it is not enough for the clothing industry which emphasizes the combination of production and marketing.
The design and production of clothing is a lasting and eternal industry (Tangchaiburana 2017). With the increasing of economic and consumption requirements, the clothing industry has rapidly developed. Consumers pay more attention to their personal preferences and the diversity of style matching when they choose clothes. In order to cater to the interests of consumers, a large number of unique and novel clothes are introduced into the market by relevant enterprises or designers. The style of clothing is developing towards diversity and personalization, and will undoubtedly become the main trend of the future development of the clothing industry. However, the complicated clothing styles and the rapid update speed make the clothing style judgment be a great burden for consumers, designers and related enterprises. Clothing industry is a special, popular, and highly professional industry. There is a common understanding in the style and various details of clothing, but it also has a strict evaluation standard. From the perspective of both designer and consumer, the interaction between the general understanding and professional standard is very important. Designers need to get the general knowledge of users in order to launch products and cater to the public. For consumers, designer’s understanding of professional standards will help them to choose more fashionable clothes. Nowadays, non-professional consumers’ understanding of clothing style is mainly through the experience from their partners or themselves. They have few opportunities to contact and understand professional standards. It has become an emerging issue to provide a platform for mutual transformation of professional standards and public cognition (Wang 2019).
Image is the main form of clothing information. For either design drawings or photos, the clothing information is transmitted through images, and various styles of clothing are intuitively reflected. From the perspective of computer, images are easy to collect, store, and transmit. Thus, image becomes the best object for clothing style research. From the professional view, all kinds of clothing styles have their relatively steady shape (Zhong 2017). If we can collect a certain number of professional and representative information of clothing styles, we can get the standards to identify different clothing styles through image feature extraction technology. The ordinary clothing image is analyzed by using image feature extraction technology (Wei 2020) to determine whether the clothing matches one of the standard styles. Then, the computer can automatically determine the clothing style to implement the conversion between intuitive cognition and professional standards. This paper first extracts the features of sample clothing images which are provided by experts. Then, the extracted features of clothing images are used to construct training set which is used to learn a classifier which can predict the style of future clothing. The main contributions are summarized as follows: