Introducing: Expiry Diary

Expiry Diary By Ryden Wiebe and Roman Gill

Project Summary: Families find it annoying to go to the household fridge and find their food products expired. Shoppers find it annoying to go to a Grocery Shop and sometimes finding products that are close to or already expired. It is a waste of money and waste of product resources. Our team was disgusted at the food waste in our city, we were inspired to think of a system that can track food and notify users (personal households or larger industry grocery companies, even restaurants) of when food is close to its expiry date or best before date. This system is not restricted to perishable food found in fridge/freezers; it can be used for any food items on the shelf that has an expiry or best before date. A list of soon to be expired food can be tracked by the user and can be used or donated to the local Food Bank. This initiative helps our hackathon theme “local issues”, which is feeding our city’s hungry people in need. Households and Companies who would have thrown away the soon to be expired food can instead donate the food to the local Food Bank, therefor contributing to the local community and not producing food wastage. Our team learned how to put a creative idea into motion. The process of having a few ideas and implementing them into a real project was a valuable learning process.

How is the Expiry Diary used? The user holds up a camera to scan a product, the system would recognize the product from the vast database and list the item and expiry date. The expiry date can be taken from the product package and auto populated in the system, or if there is no expiry listed the system can assign it a typical average “life expectancy” of that type of item. Further versions of the system could even take detailed photos of the product and determine, using past data, based on the “look” of the product an expiry date. A list is created, and as time goes on, upcoming expiry dates are flagged to the user. The user then has the option to use up the product; or alternatively, the user can select the items from the list and a message is sent to the local Food Bank to pick up the list of products. This takes the burden off the “owner” of the product, as the system will notify them that this product is about to expire, “are you going to use it? Else, I’ll get the Food Bank to pick it up.” The same for larger scale grocery companies, as close to expired products are rotated to the front of the shelves to be chosen first. Instead of being discarded, they can be donated to the local Food Bank for consumption. Hence, they do not have to dispose of the products themselves and the products are not wasted, and consumed by the local community. Win-Win for all.

Technical portion: We developed a system that uses image recognition to identify a whole database of different foods at different image angles. For example, when scanning an apple, the database would have different types and colours of apples and different angles of the apple for image recognition (side, top, bottom views), or a box of crackers (different views of the cracker box). The scanning system would also identify the expiry date or best before date stamped on the package. This would auto populate on the date field of the outcome spreadsheet. If the product does not have a date listed, the system will assign a typical average life expectancy for the product. For example, a box of crackers would have a shelf life of 9 months, the spreadsheet would assign the box with no expiry date stamp a 9 month expiry. An apple has a shelf life of 3 weeks before it start to get soft, the spreadsheet would assign the apple a 3 week expiry. The food type image would be sent to a device ran by an Arduino, it would keep track of when the products expire and alert the user beforehand (users can set parameters of how soon and how often they want to be notified i.e. 3 days or 5 days before expiry date). Once alerted the user has the option of using the food or having the system shoot off a message to the local Food Bank to pick up the products reducing food waste. This system can be used in personal households, or a larger scale grocery store, or even restaurants so far. We have made a prototype that uses teachable machine learning for image recognition so the database of product images is always learning and growing. The Ardunio based device runs a simplified program that connects with image recognition and produces a spreadsheet sorted by date (expiry). Our team’s biggest challenge for this project was to get the image program to connect to the Ardunio. We know it can be accomplished, but we did not know how to do it at our skill level (since we are only in grade 10) Please vote for our team project. We may lack the technical coding expertise, but we have big creative ideas.

Team: Ryden Wiebe and Roman Gill Eric Hamber Secondary, grade 10