A domestic research team has developed artificial intelligence (AI) technology that analyzes personal tastes and the latest trends in social media, using a database of 6 million items to help users create their own unique fashion products.

The Electronics and Telecommunications Research Institute (ETRI) announced on the 1st that it has developed an AI fashion product market platform technology that analyzes various individual media platforms to identify trends, reflects these trends in designing new clothes, and even allows for virtual fitting.

The technology developed by ETRI uses AI to create tens of thousands of new designs that reflect users' tastes and the latest trends, allowing users to virtually try on the designed clothes.

In the actual design field, if you select conditions such as the type of clothing, season, color, and pattern, AI will create a custom design.

You can try it on a model and visualize it again, so you can predict a realistic finished product before actually manufacturing it.

ETRI listed the following as the core AI technologies of this technology: △fashion video multi-information extraction technology, △new design creation and style conversion technology, and △AI technology for automatic creation of outfit videos. />
Using this technology, even ordinary people can easily simulate the planning and production of cultural products such as fashion and accessories using AI.

This is because even without design knowledge, it is now possible to select a design recommended by AI, commercialize it, and even apply it to a virtual model.

By utilizing the technology developed by ETRI, small business owners in the fashion industry and individual media creators will be able to achieve competitiveness comparable to that of branded product sales, and will provide significant assistance in consumers' purchasing decisions.

The research team explained that this technology is technologically different from the method of adding clothes using augmented reality (AR) as it creates models and videos wearing clothes according to one's intention.

He added that the research on analyzing human posture, such as home training and golf swing posture analysis, was of great help in researching automatic creation of clothing videos that take into account the model's posture.

The research team recently built a large-scale fashion dataset of over 6 million images specialized in K-fashion, and plans to present results that reflect more Korean trends in new designs and model images generated by AI.

ETRI Content Research Center Director Jeong Il-kwon said, “AI technology has moved beyond the process of analyzing and understanding data to the realm of creation, making innovative contributions to the actual product creation process and enabling the creation of new revenue models.”

This study was supported by the Ministry of Culture, Sports and Tourism and the Korea Creative Content Agency as part of the Cultural Technology Research and Development Project, and was participated by Designnoble Co., Ltd., Omnius Co., Ltd., and Sungkyunkwan University.

The large-scale data set used for AI learning was secured through the construction of AI data on fashion products and videos of them being worn, which is part of the AI learning data construction project supported by the Ministry of Science and ICT and the Korea Institute for Advancement of Intelligence Information Society.