Space Clowns

• 

  Image detection with reference object

• 

  Image detection with reference object

• 

Inspiration

We got the inspiration for this app from the struggles we face in our daily life. Vaisakh and Arnav have ADHD, and struggle with spatial management and keeping their spaces clean and organized. Suraj has OCD, and feels the need to hyper-optimize his space. While we were discussing how ironic it was that despite having the opposite problems, we end up with the same roadblock: a need for an easier way to organize our spaces.

What it does

Space Clowns is a web app that, using camera photos, can find a way of packing many objects into many shelves

How we built it

We built it by breaking it down into 3 major chunks. 1) Object identification: Finding the dimensions of the objects in the images 2) Space optimization: Finding an optimal way of fitting the objects into shelves 3) Display: Finding a way to display the obtained solution back to the user in a way that would allow for the user to pack the items accordingly. We tackled the first problem by using OpenCV. We still had the issue of needing a reference object, and only being able to get 2 dimensions accurately at a time. We decided to take user input for the reference object, and use 2 photos: a top-down view and a side view in order to get accurate 3 dimensional data. We tackled the second one by using a simple algorithm for cuboidal packing we found in a research paper about optimal packing. As packing is a NP-hard problem, finding a simple algorithm that makes it computationally efficient was a large requirement to make this packing accessible to everyone. We used python and 3d information to find positions and orientations for each object For the last problem we developed a web app using streamlit in order for the user to input their data. We created an interactive animation using plotly to demonstrate how to pack the objects.

Challenges we ran into

We ran into many challenges while creating this product. We struggled to find a good way to find the 3 dimensional data from images. We considered many methods, such as using lidar or video, or many images for different angles with all the objects at once. We finally landed on our solution of using images with a reference object.

Accomplishments that we're proud of

We are proud that we are able to develop our idea into a simple and effective tool that is practical and efficient for our target group, the people who we relate to.

What we learned

We learnt a lot about developing an app front to back. We also learnt about how to flesh out an idea, and what it takes to make a tool that is functional for the majority, not just a niche.

What's next for Space Clowns

With more time, we would love to add gamification and frequent reminders (or what we call ADHD mode) to help people with ADHD.

Built With

• opencv
• py3dbp
• python
• streamlit