Hong Kong : PolyU & Alibaba join hands for ‘FashionAI Dataset’

0
36


Courtesy: PRNewsfoto/The Hong Kong Polytechnic/ (From left): Prof. Calvin Wong, Prof. Wong Wing-tak, and Menglei Jia

The Institute of Textiles and Clothing (ITC) of The Hong Kong Polytechnic University (PolyU), and the vision and beauty team at Alibaba Group, specialising in vision intelligence and applications, are set to establish the ‘FashionAI Dataset’ for systematic analysis and labelling of fashion images based on ‘fashion attributes’ and ‘key points’ of an apparel.

By integrating fashion knowledge and machine learning formulation, the establishment of the Dataset will enable machine to better understand fashion, bringing a new horizon to the fashion retail industry through the application of AI.

Current fashion image searching technology used on online platforms is based on the whole fashion image to search the exact or other similar images. However, if a customer is interested in some particular fashion attributes of a fashion image and wants to search other fashion items with these attributes, the current searching technology cannot meet the needs of the customer. This greatly limits the potential development and applications for offering more customised shopping experience.

From artificial intelligence (AI) research perspective, this limitation of the current image searching technology is caused by the absence of available fashion image dataset constructed with both fashion professional knowledge and fulfils the requirement of deep learning, that is, the current technology is unable to train a machine to accurately understand and recognise the fashion attributes of each fashion image.

Fostering the application of AI in the fashion industry, a PolyU research team led by Professor Wong, worked closely with Alibaba to develop ‘FashionAI Dataset’ to solve two fundamental problems of the deep learning algorithm; ‘apparel key points detection’ and ‘attribute recognition’.

Key points (for example, neckline, cuff, waistline) and fashion attributes (for example, sleeve length, collar type, skirt style) build the foundation for machine learning in understanding fashion images. The establishment of key points and fashion attribute database enables the computer to effectively and efficiently understand the fashion image which is fundamental for deep learning and recognition algorithms.

The accuracy of key points detection is determined by several factors such as the dimension and shape of the apparel, distance and angle of shooting, or even how the apparel is displayed or the model is posing in a photo. These factors can lead to poor key points detection and result in an inaccurate analysis of fashion images by the computer. Accurate key points detection, can therefore, improve the performance of deep learning algorithms.

Fashion attributes are the basic design elements of an apparel, and their combination determines the product category and styles of a fashion item. With the wide variety of fashion attributes, attribute recognition is a complicated process. A systemic classification of fashion attributes is essential to accurately label fashion attributes, facilitating research on deep learning and algorithm design for fashion image searching, navigating tagging, and mix-and-match ideas, etc.

The revolutionary Dataset can greatly facilitate understanding fashion images and related algorithm design, and developing machine learning. It would help improve the accuracy of online fashion image searching, enhance effectiveness of cross-selling and up-selling, create innovative buying experience and facilitate customisation of online shopping platforms. (GK)

Fibre2Fashion News Desk – India



Source link

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.