ORA is a food optimization concept created to target food wastage within the public domain. Hinged on emerging technologies such as LiDAR and Ai inference on the cloud, we have established a fully integrated diet tracking and meal recommendation system for customers and restaurants alike.
In the public dining sector, consumers frequently fail to gauge meal portions based solely on photos, while restaurants struggle to collect feedback on meals that are left incomplete. These wastage issues have been noted to exist on a global dining scale and are further accentuated when side dishes are added to the equation. Ora aims to target this complex issue at every level.
The Ora system is comprised of three intertwined products. These are the customer app, restaurant monitoring software and finally the camera mount. These 3 factors are bound by an overarching campaign that is designed for appeal and education of current sustainability issues and more importantly, global adoption in the public domain.
The first of the three deliverables is the Ora Mobile application. The app is designed for optimised food ordering. Customers can easily browse restaurants that use the Ora system and select the intended size of the meal. In addition to small, medium and large, Ai is used to recommend a size based on previous eating habits. Side dishes can also be specifically ordered based on the customer’s taste. After eating all the food, customers are incentivised to take photos of their clean plate to collect AR digital collectables.
Inclusive in nature, our solution is built from the ground up to procedurally adapt data from a wide variety of global cuisines through a simple photograph and reconstruct the image in 3D. A neural network is used to identify ingredients while LiDAR gauges the depth and quantity of the meal - both at the ordering and meal completion stage. To help restaurants collect this data, we have created a ceiling-mounted LiDAR camera installation specifically optimized to function in a kitchen.
When meals are prepared and empty dishes are returned, the plate is passively photographed from a ceiling-mounted camera. While any commercial camera can be used, we proposed the use of a LiDAR embedded unit, as it will allow our Ai to accurately extract a viable height map from the dish before and after consumption.
We asked the question: How might we improve the public dining experience to reduce food waste to meet tomorrow's sustainability goals? Developed for a post-Covid restaurant context, we took a macro-level approach to revise the customer dining experience while taking data collection opportunities into account.
Our first point of contact is the ORA app. Customers order, personalise or select recommended meal parameters from their favourite restaurants. The size and contents of the recommended meal are dependent on consumption history.
When the meal is prepared and ready to be delivered to the customer, The restaurant's RGB depth camera registers a photo of the meal - which is then procedurally recreated in a 3D simulation using an object classification process to assign 3D models to the segmented ingredients of the photo.
Users are incentivised to complete their meals through our rewards system. The ORA app allows users to take a photo of their empty plate after a meal - if the plate is detected to be empty a rare digital collectable is unlocked. This could include, coupons, stickers or social media filters.
Our process for data collection hinges on a trained neural network that is optimised for meal detection and ingredient classification. Dishes are scanned before and after a meal for digital re-creation in 3d and assigned to a customer's profile for autonomous analysis on eating patterns. Our algorithm is responsible for sorting this data into trends and offering healthy suggestions - which manifests as the recommendation feature within our app.
Inclusive in nature, our solution is built from the ground up to procedurally adapt data from a wide variety of global cuisines through a simple photograph and reconstruct the image in 3D. A neural network is used to identify ingredients while LiDAR gauges the depth and quantity of the meal - both at the ordering and meal completion stage. Through the use of monitoring and pattern recognising algorithms, each users meal habits are recorded and used to personalise their meals factors. A procedural function was created in Blender 2.93 to achieve our effect - an example node set-up is recorded below.
Ora’s slogan is “Clean plate, Clean planet.” Proliferated through social platforms such as Instagram, we use analogue methods such as posters in public areas to encourage the adoption of our campaign and there-by the importance of sustainable food habits. Using fun digital experiences, we improve our relationship with restaurant owners and customers in the public domain while encouraging users to reduce food waste at large - in this way, we emphasize how simple changes can make a huge difference.