Garde is an automated fencing coach being created with Davis Fencing Academy, a Team USA Fencing certified association. Garde aims to perform a robust analysis of a fencer's form and harness generative A.I. to give detailed feedback as to how the fencer can gradually improve over time. Additionally, Garde allows coaches to access their students' videos and give individual feedback. We hope that Garde is used by leading fencing institutions not only in Davis and The Bay Area, but also worldwide. This application is truly one of a kind and is patent pending.
Currently, Garde uses Next.JS and Tailwind CSS to create an intuitive and immersive frontend, alongside Tensorflow.JS, and Plotly for pose estimation and visualization. User authentication has been implemented in house to fit our infrastructure, and offers a seamless transition between the landing page and the fencer and coach pages. For the backend, Garde uses Cloudflare D1 and Cloudflare R2 to store valuable user data and videos. Garde uses ML to analyze and visualize fencer form, and uses generative AI to give valuable feedback that the fencer can use.
The landing page for Garde. This page gives users all of the information they need about what Garde does before they decide to use it. Also, this page has a contact form so that coaches who are interested in using the product can contact us for access or inquiries.
The fencer page for Garde, this page is used by fencers to perform the instructions shown above the canvas and get the feedback they need to improve. Garde calculates the angles between their joints, speed and distance in real-time to offer them key information they can use to assess their performance.
The coach page for Garde allows the coach to see all thier students' videos and give them feedback. Once the coaches are done seeing the students' performance, they can also view their progression over time which is detailed through graphs.
The tournaments page for Garde allows users to manage and organize tournaments, it allows for participant signups and sharing essential information like location, schedule, registration fees, organizer information, participant profiles, descriptions, rules, and eligibility criteria. It also allows easy access of tournament videos and results.
Currently, Garde is patent pending and the code is private. However, if you are curious about Garde and would like to know more, please contact support@gardeai.com.