CORDIS - EU research results

Virtual reality fitting simulation for electronic e-commerce

Periodic Reporting for period 1 - SP3D (Virtual reality fitting simulation for electronic e-commerce)

Reporting period: 2016-06-01 to 2016-09-30

Online fashion industry worldwide losses Billions of dollars on returned goods; 62% of all deliveries are wasteful. The cost of returned goods includes not just the freight cost but also pays its toll on customer satisfaction, customer loyalty, brand perception, marketing resources, and environmental contamination increasing unnecessary waste and pollution. Reason enough to state that e-commerce processes are far from efficient cutting down retailers’ benefits over 25%.
Today’s fashion e-commerce is asking the customer to follow their sixth sense to guess if a piece of garment at the online store will fit them. Same reason people don´t get into a physical store to buy clothes without trying-on them first; e-shoppers are reluctant to buy clothes online and in result only 26% of people with access to internet dare to buy fashion products online. What online shopping experience offers, although convenient, it is far from the traditional shopping experience and even farther from its true potential.
The main objective of the overall innovation project is to bring the latest innovations in 3D, computer vision and virtual reality to make possible for people to try-on clothes in their mobile devices like if they were looking themselves at the mirror.
ShapeMe have explored the possibility to use time-of-flight sensors to obtain automatically HD quality 3D models, in a controlled static environment using multiple sensors. The main goal for this “phase I” has been to assess the feasibility of the technology applied to the apparel industry, looking into both the current state of art and the need for further development to reach commercial viability and finally to deliver a complete business plan.
Time of flight sensors are inexpensive in comparison to 3D scanners which make them ideal for commercial applications. On the other hand, these sensors face several downsides that condition the usability of the technology and affects both the setup and the working environment.
Throw this feasibility study we have found the optimal working setup, environment and processes to adapt the technology in order to obtain 3D models in HD quality. The use of apparel models in a virtual reality environment will simulate the fitting in a realistic manner and thus shoppers will be able to see the fit on their mobile devices without going throw the actual fitting room.
The R&D team faced several challenges regarding the sensors technology and its need for accurate calibration. The amount of work that needs to be done on each sensor to achieve a proper calibration is considerably large. RGB cameras and IR sensors needs to be calibrated individually and then go throw a stereo calibration to properly match them. This calibration gets even more challenging when multiple sensors are used such is our case with a setup of up to twenty sensors. The most challenging of all is what we have called “Z calibration” since each sensor will provide a different value in every single pixel for every single range not following a linear progression which makes it extremely challenging to achieve what we set as the adequate accuracy <2mm.
Besides the technical and technological aspects of the project we have conducted a market analysis -to identify the commercial potential of the idea; understand client needs; define the competitive environment-an internal analysis to better understand our capabilities; need to partner; need to reinforce the team- fine tune the business model, sales forecast, balance sheet, cash-flow, asset the financial viability of the project and define an implementation plan setting up road map and finally a marketing plan.
Being able to generate HD quality 3D models automatically is already a big step but doing so in a cost efficient way will open a path for multiple commercial applications. 3D scanners are costly and mostly use for industrial purposes. There is no standard commercial software to calibrate and process sensors and their data. Usually for each application and hardware there is the need for a specific software. To achieve our innovative idea, we need to develop specific software in the fields of computer vision, point cloud reconstruction and virtual reality to create the right working tools that allow for making a realistic fitting simulation.

Achieving so will have a big impact on e-commerce, cutting down losses and wasteful deliveries, diminishing pollution and waste and creating a disruptive user experience. Shapeme will open the road for new ways of doing businesses. Shapeme will build a unique data base with millions of body measures that could be used for numerous purposes and analysis, medical studies and so forth. For instance, it could be used to standardize sizes globally or track body changes throw someone’s life or use it to tailor made products and experiences.