Variety Labeler for Image Recognition

A prototype was developed for accurately labeling different varieties of Peruvian potatoes from digital images. Users can easily add labels for four native potato types (Peruanita, Amarilla del Centro, Huayro Rojo, and Huamantanga), with each label distinguished by a unique color. The labeler automatically generates bounding boxes around the potatoes, which can be adjusted for precise fitting. Once labeling is completed, the information is saved within the image, and the total number of bounding boxes is displayed for each potato photo in the system. The labeler's architecture comprises a frontend implemented in Vite React, deployed on Vercel, and connected to the backend through Axios. The backend is built with Node, deployed on Heroku, and incorporates New Relic and Sequelize ORM for efficient operations. It is connected to a PostgreSQL database for data persistence. Additionally, the information is stored in AWS.