SignWave

Meet Budgeting App Type Beat

Anas Almasri is a grade 12 student at John Fraser Secondary School, who is currently interning at Bassem Labs. His hobbies are geography, history, and soccer. He codes in Python, JavaScript, and React. Some fun projects he previously created are a Flag Quiz Website and a UCL Draw Simulator. His LinkedIn, DevPost, and GitHub.

Anas Alsaid Ahmad is a first year chemical engineering student at the University of Waterloo. His hobbies include swimming and skiing. He codes in Python, Rust, and TypeScript. His favourite project before HawkHacks was a Connect4 game he built with a custom AI opponent. His LinkedIn, and GitHub. He does not have a DevPost.

Marwan Al Kharrat is a grade 12 student at Stephen Lewis Secondary School. There, he leads the hack club, where he tries to help others expand their portfolios. He is experienced in Python, Java, and C++, and loves to make Telegram Bots. In his free time he likes to watch movies, write, and release music under the pseudonym ATHEMOSU. His LinkedIn, DevPost, and GitHub.

Mohammad Alwattar is a grade 12 student at Bishop Reding Catholic Secondary School. In his free time he reads about mathematics and physics, and works on learning different programming languages. As of now, he has experience in C++, Java, Python, Rust, and Swift, but he is yet to decide on his preferred one. He was part of the team that won the Best UI/UX category in Ignition Hacks 2023 where his team coded RecipeLens. His LinkedIn, DevPost, and GitHub.

What is SignWave?

SignWave is an educational app that teaches its users sign language. It starts off by teaching users 5 basic phrases, and then tests their knowledge via multiple choice questions. The question base is not very big at the moment, but the code allows for very easy expansion.

Why did we create SignWave?

As the point of a hackathon is to challenge oneself, we wanted to try something new. Visual machine learning seemed interesting, and an app that could teach sign language, and then test user while being able to identify whether or not they were signing correctly seemed like a promising idea. However, due to issues out of our control with the panda library in python, we were unable to implement visual object recognition in time.

How did we create SignWave?

SignWave was made mostly in JSX, CSS, HTML, and Python. The scrapped visual sign recognition component was built using TensorFlow and Panda, however, issues with the Panda library installation led to it being scrapped for this stage of SignWave. This is something we hope to achieve in the future, especially as we got frustratingly close to success.

Difficulties faced

At this current stage, our database of sign language phrases is extremely limited, owing to the nature of hackathon projects. However, the program is easily scalable, and as we collect more data, more questions can be added. . It was also extremely difficult to set up, label, and train the AI model because of issues with setting up the Python environment, so we ultimately sidelined that component, and decided to implement it later.

GitHub Repo