Shirt Version Intelligent Recommendation for Rapid Garment Customization

This paper discusses the method for shirt version intelligent recommendation through combining human body data analysis and apparel loose quantity setting analysis. Firstly, we get the weight of each factor by Analytic Hierarchy Process (AHP) and realize the recommendation of the best shirt shape through the garment size. Then by using the Back Propagation(BP) neural network, we input net body size and output corresponding garment size for supervised learning. Meanwhile, we get the weight of each factor after linear fitting and can calculate each part of the loose quantity through the net body. Finally, we check whether the recommended shape is fit through loose quantity fitness analysis. The version recommendation in this paper not only helps customers to find fit size, but also helps the manufacturers to calculate each part of the loose setting.

I. Introduction

Currently, in the production of garment customization, the most important thing is that how to get the garment size whose version matches the company itself through the customer's net body size and choose the most approximate version to change version accordingly. Generally, customers refer to the shape on clothes when they select them. Loose quantity is the difference between the garment size and net body size. What's more, loose quantity has an important effect on the clothing comfort, style and effect.