Amabam

Teck Stack
Selecting Skills
Skills Tree
Lesson Example

Inspiration

We're interested in creating an app that teaches hobbies because we believe that hobbies can bring a lot of joy, fulfillment, and personal growth to people's lives. However, we know that getting started with a new hobby can be a challenge due to various barriers. That's why we want to create an accessible and user-friendly app that makes it easy for people to learn and explore hobbies within their own budget and schedule.

What it does

It is a gamified learning tool to learn several hobbies such as juggling. Individual skills are broken down into subcomponents, and users can choose which area they want to focus on.

How we built it

We used mongodb with flask as the backend, react for the front end. We also used mediapipe for pose recognition, opencv for image filtering and CNNs for temporal predictions.

Challenges we ran into

We had a lot of difficulty creating a suitable model that recognised good behaviour. We evantually setlted on a self-trained CNN network that was fed pose coordinates from the user. This was due to its proven record in temporal predictions and its relatively light weight nature.

Accomplishments that we're proud of

First off, we created a fully usable user log in site with encrypted passwords.

We managed to gather data, and build a working model for computer vision aided juggling. We also created a gamified, skill-tree based website that breaks down learning into simple components.

What we learned

That integration is hell.

What's next for Amabam

We have been experimenting with zero-shot learning so that users may generate content that they require for themselves. It has shown some success so far.