Inspiration
Creating user interfaces doesn't come easy to everyone. We wanted to make prototyping designs quick and easy. Optical character recognition was one of the major features we wanted to implement into our program.
What it does
- Handwritten or computer generated text is analyzed
- Convert analysis to a jsx file that is rendered
- Redo analysis whenever there is a change
How we built it
- EasyOCR reads handwritten and computer generated text
- Our backend uses a Node server to receive JSONs
- The backend creates jsx files
Challenges we ran into
- When trying to implement the OCR functions of several cloud services (Azure, GCP, AWS), we ran into issues with credentials
- Recognizing hand written text (balancing latency vs accuracy)
- Sending images via POST
- Integrating all the components
Accomplishments that we're proud of
- Program can recognize text
- The backend server is capable of writing a jsx file
What we learned
- OCR training
- Writing to a file with escape characters using Node
What's next for Gen Dev
- Send images directly via POST
- Improve file management system
- Using vector calculations to analyze whiteboard movement

Log in or sign up for Devpost to join the conversation.