TBOER

• 

  The user taking pictures of their cube to be solved.

• 

  The animation of the cube being solved being played out for the user to follow.

Inspiration

The pain of not being able to solve a Rubik's cube affects roughly 180 million people worldwide(No, this statistic was not purely made up). It might not be the most intense issue facing humanity in the modern age, but it certainly is both extremely widespread and very annoying.

What it does

The goal was for it to do this: Given user-provided pictures of faces of their unsolved rubik's cubes, it uses machine learning to analyze and identify the colors on each of the squares and then shows the user the moves they need to perform on the cube to solve it.

How we built it

We each worked on individual parts of the project, and at the end tried to consolidate them into one product.

Challenges we ran into

We are guilty of procrastination, and much of the project was done during the final stress-filled four hours. We also had many issues with getting the animations to work, hence the plethora of goofy cube footage within the video presentation.

Accomplishments that we're proud of

We're proud that we were each able to uphold our ends of the product, and come together at the end to have a mostly complete app despite its overall enormous complexity.

What we learned

We learned an enormous amount about teamwork and time management(and the importance of sleep!). We also were also enlightened to new animation techniques, the wielding of AI, and the need for good training data.

What's next for TBOER

The project is not finished at the moment, but it is roughly 70-80% complete. The final step would be to link the cube solving algorithm that the app generates with the animations. Of course, training the AI with more quality data would be helpful. Detailing the animations would also help the user more clearly see what they need to do to solve the cube.

Built With

• javascript
• machine-learning