Filterai

• 

  Architecture Diagram

• 

  User Flow Diagram

Inspiration

E-commerce has greatly evolved the way we shop, and there have been more businesses online selling clothes than ever. A problem many retailers might face is they do not know the exact size when shopping for clothing items, or even after measurements, they might not get the correct size they want. For business owners to deal with such problems, hinders their store's efficiency and could potentially incur more costs to them.

Our solution aims to be a highly accurate and realistic Augmented Reality (AR) sizing solution. This solution would allow shoppers shopping online on Zalora’s platform to be able to test and 'try on' products that they are interested in from the comfort of their own homes, thereby reducing any need for swaps and exchanges with the business. Another benefit is business owners might be able to garner additional data-driven insights by using our AR solution.

What it does

Our solution is an API that fashion commerce companies can easily implement in order to generate Augmented Reality models of the clothes they are selling to the user. Users would be able to turn on their cameras and get a live feed of the clothing model being overlaid on their own bodies. The user could also opt to upload a photo instead, and our API would be able to overlay the size that they selected onto their body.

Besides Users, businesses can also use data on how much time or retention a particular model or type of clothing might enjoy. These analytics can contribute to further data on what type of clothes users like, what type of clothes might be more effective to advertise with an AR model etc. Analytics-driven insights from using the API could potentially increase a business’s model, and all they need to do is include our API and integrate their store with 3d models of their clothing.

The following is a possible user flow for how retail shoppers could potentially interact with our API when they are shopping for clothing.

How we built it

We made use of AWS infrastructure in order to design the architecture required for our application.

AWS API Gateway: The AWS API gateway would be used as an API management service for the AR application to do RESTful API requests to AR components that would deliver the AR content to the users. The use of a gateway would create a seamless connection from the frontend services to the AR pipeline, reducing latency and improving the AR experience for the user.

AWS Cloudfront: One of the main challenges observed in our proposal would be to deliver large 3D assets to fit and test clothes via the application in real-time (AR streaming pipeline). This real-time application requires low latency in order to create a seamless experience for the user (real-time update of 3D according to the user's camera position).

Therefore, AWS Cloudfront can be used as a content delivery network (CDN) to stream and deliver 3D assets to users in real time. With its globally dispersed Points of Presence (PoPs), it would reduce the latency of content delivery across the globe, providing the Zalora application the possibility to stream its AR content seamlessly to users on a global scale.

AWS Lambda: The Lambda service provides an event-driven computing layer that would provide on-demand computing of AR calculations such as the clothing simulation and rendering of 3D assets to the users. Using this event-driven pipeline to the architecture would allow for processing of the AR service to users at scale without the need for provisioning for server resources, allowing for higher utilization of resources.

AWS Cognito: The AWS Cognito allows users of the Zalora app to sign up using credentials from popular social media networks (SSO), this reduces the barrier of entry for users who want to use the app without the hassle of signing up for a Zalora account.

Team

Built With

• amazon-web-services
• aws-api-gateway
• aws-cloudfront
• aws-cognito
• aws-lambda