PolyU and Alibaba cooperate to launch FashionAI Dataset
28 August 2023
In order to strengthen the application of artificial intelligence in the fashion industry, Hong Kong Polytechnic University (PolyU) and e-commerce giant Alibaba have jointly developed the world's first FashionAI dataset to better meet the needs of fashion retailers and consumers.
At the recent Taiwan-Hong Kong Economic Cooperation Forum in Taipei, Wong said that by integrating fashion knowledge and machine learning methods, data sets will enable machines to better understand fashion and open up new horizons in the fashion retail industry through the use of AI.
The data set can solve two basic problems of the deep learning algorithm: the detection of clothing key points and the identification of clothing attributes. Key points to be tested include the neckline, cuffs and waistline; and the attributes to be identified include sleeve length, collar and skirt. Wong said that both lay the foundation for machine learning to understand the fashion image.
He said that Dataset can help improve the accuracy of online fashion image searches, create innovative buying experiences, and facilitate customization of online shopping platforms.
E-commerce is now the main market for the fashion industry, and many fashion brands have online sales of more than 50%, Wang said, adding that some cosmetic brands have applied artificial intelligence-based facial recognition technology to provide virtual makeup testing services.
In addition to using shopping habits and preferences, Wong said, e-commerce operators are also looking to better identify the size and size of their customers in order to increase promotions and reduce the return rate of sales apparel.
To meet this demand, Wong revealed that PolyU has developed an intelligent 3D human modeling system by integrating cutting-edge computer graphics and vision technology. The system allows for the creation of custom models of any dynamic pose automatically within 5-10 seconds, enabling the detection of body parts in tight-fitting garments from front and side-view 2D images and then building them into 3D models.
In general, the birth of the FashionAI data set, let us focus on the basic issues of machine cognition fashion, together with the promotion of AI technology in the fashion industry.
At the recent Taiwan-Hong Kong Economic Cooperation Forum in Taipei, Wong said that by integrating fashion knowledge and machine learning methods, data sets will enable machines to better understand fashion and open up new horizons in the fashion retail industry through the use of AI.
The data set can solve two basic problems of the deep learning algorithm: the detection of clothing key points and the identification of clothing attributes. Key points to be tested include the neckline, cuffs and waistline; and the attributes to be identified include sleeve length, collar and skirt. Wong said that both lay the foundation for machine learning to understand the fashion image.
He said that Dataset can help improve the accuracy of online fashion image searches, create innovative buying experiences, and facilitate customization of online shopping platforms.
E-commerce is now the main market for the fashion industry, and many fashion brands have online sales of more than 50%, Wang said, adding that some cosmetic brands have applied artificial intelligence-based facial recognition technology to provide virtual makeup testing services.
In addition to using shopping habits and preferences, Wong said, e-commerce operators are also looking to better identify the size and size of their customers in order to increase promotions and reduce the return rate of sales apparel.
To meet this demand, Wong revealed that PolyU has developed an intelligent 3D human modeling system by integrating cutting-edge computer graphics and vision technology. The system allows for the creation of custom models of any dynamic pose automatically within 5-10 seconds, enabling the detection of body parts in tight-fitting garments from front and side-view 2D images and then building them into 3D models.
In general, the birth of the FashionAI data set, let us focus on the basic issues of machine cognition fashion, together with the promotion of AI technology in the fashion industry.
