This repository contains the source code and development tools for Mull Recognition, which is a standalone application implementing the waste detection tool from RGPosadas/Mull.
Mull Recognition implements real-time object detection in the browser, and is capable of classifying the appropriate action to take for the given waste (trash, recycle, etc.). It is implemented using Tensorflow, and leverages a TensorflowJS model trained on custom data.
See cristian-aldea/mull-model for the model files used by this application.
To develop and contribute to this repository you will need the following:
To install the project's dependencies:
npm i
To serve the application locally:
npm start
To lint the typescript code using eslint:
npm run lint
To upgrade all packages to their latest version:
npm run update
See the the docker README for details on how to build and run application for production.