DishNomo | Devpost

**Do YOU live with others?

Do YOU consume food?

Do YOU have issues with others not washing their dishes?**

Congratulations! You're 100% eligible for our life-changing solution, DishNomo, as inspired by the average university roommate experience - too many dirty dishes in the sink!

So what does it do?

Using facial recognition, our patented invention will remember who is actually leaving their dishes in the sink. If they leave their dishes in the sink a little too often and wash them a lot less, our friendly addition in the cabinets will LOCK it up, until they clean their dish. If you can't clean 'em, why use 'em?

How we built it

Using OpenCV, we leveraged its object and facial recognition features to encode facial features, identify coloured plates, store movement, and obviously, recognize the perpetrators leaving all those bowls behind. Towards the hardware side, we used an ESP8266 WiFi board to control the servo that acts as the locking mechanism in the cabinet (inside, so it's not an eyesore). Via wifi, we can send instructions to the ESP8266 board on how to control the servo.

Challenges we ran into

Getting OpenCV to work was definitely the hardest problem. We frequently got the dreaded "Import could not be resolved error", which resulted in many hours of uninstalling and reinstalling. Finetuning the program to actually register the objects/faces we wanted to detect was another challenge, on top of all the finicky hardware problems. We didn't expect it to be a smooth ride, but we didn't expect it to be this bumpy either. From square one, we struggled with setting up the right environment for our RaspberryPis, Jetson Nanos, and countless other 'simple' setups that weren't so simple after all.

Accomplishments that we're proud of

From the very start, we knew that we wanted to address a universal problem in an effective manner. We debated for hours on end about reward versus punishment, and whether we should introduce behavioural economics, before we settled on a tangible consequence to your actions. Working in a team of two provided a lot of bonding moments as we debugged for hours on end with no end in sight. We're proud to display a product that has a purpose in everyone's life - not just the broke university students. Small conflicts like dishes come up all the time, so we're glad to provide a solution that isn't fueled purely by frustration, but signals instead.

What we learned

Delving deeper into OpenCV allowed us to understand and realize that image recognition isn't intimidating after all. It's the setup that you should be scared of. As for the hardware components of this project, working with wifi modules and precise motion control showed us the inner workings of the tiny tech around us.

What's next for DishNomo

With more time, money, and materials, anything is possible! Specifically:

Integrating the scripts onto portable microcontrollers such as RaspberryPis or Jetson Nanos, which would be able to handle the computation
Develop the object recognition software to be adaptable to different dishes, cutlery, and pans
Integrate a leaderboard display to show users what their dish status is
Develop a system to incentivize the dish washing process - food incentives, charity chip in, purchase of choice
3D-printed casing and attachments for motorized cabinet locks (they're not meant to be strong enough to withhold the Secret Service, just enough to remind your roommate that they have dishes to wash)
Sensors, wifi modules, and microcontrollers that are better suited for the purpose of close-range communication (ie. RF transmitters and receivers are faster than standard Bluetooth modules)