Smart Fitting Mirror

An exciting technical revolution in retail industry, designed to bring the convenience and speed of the online shopping experience into physical store, enhancing retailers' connection with existing customers while attracting new ones.
Introduction
GIGABYTE's Smart Fitting Mirror is an O2O (Online to Offline) retail solutions, featuring an integrated IoT hardware with 3D ToF camera and 3D Virtual Try-On software, for users to generate virtual personal image to have dressing simulation and increase interaction with retailers.
As the interaction records increase, the data are transmitted and stored in GIGABYTE's Cloud storage system for big data analysis through AI learning, letting retailers study buying behaviors and preferences to deliver advise in real-time or provide favorable products to gain customers' loyalty.
User Scenario
On-Line Shopping
The image data is stored on the Cloud. Users can shop and try on clothes with virtual personal image and do outfit comparisons on APP at anytime anywhere until they find the perfect ones to complete orders and have them delivered to assigned address.
Off-Line Shopping
Shoppers can feel the texture of clothes in retail stores and have outfit simulation on Smart Fitting Mirror (or mobile APP) to avoid long waiting line in the fitting room. Real-time inventory information such as sizes, colors or similar patterns are also shown on screen for customers to choose.
Challenges in Retail Market
On-line shopping becomes popular to consumers not only for its efficiency but also the convenience it brings.  Physical retail stores are losing customers and facing problems of how to draw customers' attentions back to their shops.  With limited manpower in the shop, retailers are unable to have intensive interaction with customers to provide clothes wear advice, let alone catching their fashion preferences correctly to have well inventory management of various garments and accessories.  On the other hand, waiting in the long queue at the fitting room on the weekend or holidays is also a nightmare to customers.  Coming back and forth between aisles and fitting room to find the right size of outfit is also frustrating.   
IoT technology and big data analysis lead the way to improve customer experience and create return on investment by connecting customers and retail stores online and offline.
Benefits
suitable to Reduce Expenses, Money Saving, Reduces Cost
Cost Saving
With interaction results sent to cloud for analysis, retailers predict fashion preferences to adjust various stocks of clothes, to minimize costs.
suitable to Time saving
Time Saving
Users try on clothes virtually through APP and feel the texture in the store to avoid long waiting line in the fitting room.
suitable to Quality, Accuracy
Precise Marketing
With the combination of network data collection and analysis, stores provide favorable services to customers more accurately.
suitable to User Friendly, Ease of Use& Lower maintenance requirement
Causing Community Power
Shoppers can share outfit combinations on social media from the mirror to cause extensive discussion on the market.
Smart Solution with Hardware/Software Integration
GIGABYTE uses industrial-grade mainframe mITX-H31CA motherboard equipped with 8th generation Intel® Core™ i processor, a 14nm CPU which features power efficiency and support dual channel DDR4 memory bringing cutting edge features and ultimate performance. In addition, cooperating with ULSee's U'style AR/VR software solution, GIGABYTE's high-end graphic card powered by Nvidia GeForce RTX™ 2080 GPU integrated with 8GB GDDR6 256-bit memory and GIGABYTE's new innovated 3D ToF camera which is a low-cost smart solution with novel 3D imaging capture technology to bring vivid virtual personalized image on AUO 42" high-resolution touch screen (P420HVM02.0) for users to start the interactive journey.

With continuous interaction of shoppers and the system, data like outfit combinations or purchase records are sent to back-end GIGABYTE server (Deep Learning appliance: G481-HA1, G291-Z20) working with the on-premise AI training program developed by ITRI, to provide big data management & analysis of shoppers' preference to bring best choices for customers to choose.
Features of ULSee's U'Style 3D Virtual Try-On Solution
3D Digitization
The foundation of ULSee’s U’Style 3D Virtual Try-on Solution is digitizing of real garments as 3D model assets. ULSee’s in-house-developed 3D modeling solution supports processing photos for generation of high-quality photo-realistic 3D models. The 3D modeling solution can also support auto-background removal, picture location calibration as well as formation of 3D meshes and high-resolution texture maps. 15 minutes of time is all that is required for 3D digitization of garments.

Multi-Layer Garment Styling
U’Style apps support multi-layer styling of garments. By marking layer indices, U’Style apps can render garments from the innermost layer to the outermost one, thereby supporting layering of multiple types of garments of inner, top, outer and bottom types.

Customizable Personalized 3D Avatars
In U’Style apps, users can get customizable 3D personalized avatars with only one full-frontal-view photo of their faces. The customizable personalized avatars can then be used for 3D Virtual Try-on. This feature allows users to have a more-intimate experiences in interacting with U’Style apps. In addition, it is also possible for users’ customizable personalized avatars to be further-customized with hairstyle changes and eyewear virtual try-on. Therefore, customization of personalized avatars is not only a value-adding feature but is also a fun-element-introducer for an enhanced user experience. Users can get entertained and share with friends.
Products
1/4
mITX-H31CA
MINI-ITX motherboard
2/4
GeForce RTX™ 2080 WINDFORCE 8G
GV-N2080WF3-8GC (rev. )
3/4
Server G481-HA1
G481-HA1 (rev. 100)
4/4
Server G291-Z20
G291-Z20 (rev. 100)
Related Technologies
虛擬機器
虛擬機器(Virtual Machine)是一種軟體電腦,可以像實體電腦一樣執行作業系統和應用程式。虛擬機器由一組規格和組態檔組成,並由主機實體資源支援。每台虛擬機器都擁有可提供與實體硬體功能相同的虛擬裝置,這些裝置更易於攜帶、管理,且更加安全。
邊緣運算
邊緣運算是一種網路運算架構,運算過程盡可能靠近資料來源以減少延遲和頻寬使用。目的是減少集中遠端位置(即"雲")中執行的運算量,從而最大限度地減少異地用戶端和伺服器之間必須發生的通信量。近年來,技術的快速發展使硬體趨向小型化、高密度以及軟體的虛擬化,讓邊緣運算的實用度更加可行。
IoT
IoT是網際網路、傳統電信網等資訊承載體,讓所有能行使獨立功能的普通物體實現互聯互通的網路。