Orbit

Inspiration

The inspiration for Orbit emerged from a shared belief in the power of cultural exchange to foster understanding and unity. In an increasingly interconnected world, it’s easy to remain siloed within familiar cultures and experiences. Dance, however, is a universal form of expression that transcends language and geographical boundaries. It carries rich histories, unique traditions, and profound meaning within different societies. We recognized that by engaging people in learning and participating in the dances of other cultures, we could build a deeper appreciation for the art forms that make each culture distinct. Our goal was to create an interactive and immersive tool that not only showcased these dances but also allowed users to engage with them, fostering a sense of global community and respect for cultural heritage.

What it does

Orbit is an interactive app that introduces users to dances from various cultures and allows them to practice these dances in real-time. Users select a dance and replicate the moves in real time. The app uses computer vision to analyze the user’s performance, providing an instant feedback score based on how closely the user’s movements match the original dance. Each dance is also accompanied by valuable cultural information about a dance's origins and significance. This scoring system helps users improve their technique and deepen their understanding of cultural dance forms, making learning both educational and enjoyable.

How we built it

We developed Orbit using a combination of modern web technologies and machine learning frameworks. The front-end was built with React, styled with Tailwind CSS, and integrated with Firebase for seamless user authentication and real-time data management. To power the real-time analysis and scoring system, we utilized MediaPipe for body tracking and custom-trained pose similarity models using Intel AI Timber. These models were essential for providing accurate movement analysis. The app’s back-end incorporated TypeScript for reliable code management and Python for training the machine learning models, ensuring smooth operation and responsiveness.

Challenges we ran into

One of the most complex challenges we faced was calculating the intricate angles necessary to generate the target score for each dance. Ensuring that the pose similarity model accurately captured and assessed these angles across different body types and movement styles required significant refinement. Building an algorithm that could handle variations in user positioning and maintain consistent scoring accuracy was technically demanding. Additionally, integrating these angle calculations seamlessly with the computer vision system while maintaining real-time feedback and low latency posed another layer of complexity. These challenges required meticulous testing and optimization to create an effective scoring system that users could trust and learn from.

Accomplishments that we're proud of

We are particularly proud of the custom-trained pose similarity models that accurately evaluate dance performances in real-time. Successfully integrating MediaPipe with Intel AI Timber to create a robust scoring system was a major technical achievement. The intuitive user interface, which adapts well to a variety of devices, is another accomplishment that ensures the app’s accessibility.

What we learned

Developing Orbit taught us the importance of blending different technologies to achieve a cohesive user experience. We gained valuable insights into optimizing machine learning models for real-time application without sacrificing accuracy. Working with cross-functional technologies like React, Firebase, and MediaPipe expanded our expertise in handling complex integrations. We also learned the significance of user feedback in refining our app to ensure it is user-friendly and educational.

What's next for Orbit

Moving forward, we plan to expand Orbit by adding more dances and cultural traditions to our library. We also intend to incorporate social features, such as sharing scores, challenging friends, and participating in global dance competitions. Personalizing AI feedback to cater to individual learning paces is on our development roadmap, as is collaborating with professional dance instructors for expert tutorials. Our ultimate vision for Orbit is to build a thriving community that celebrates cultural diversity through dance.