/BabyVision-Backend

Primary LanguageJavaScriptMIT LicenseMIT

BabyVision-Backend

Routes (Testing) collection on postman - https://www.getpostman.com/collections/661c6bec9516dc211dd6

This code sample is basically the backend (server side code) of the Android Native Application Baby Vision which is a ML integrated application which lets users( especially children) to identify objects ( label classification) and landmarks( landmark classification) near their vicinity. This entire server side code is divided into three parts:

  1. User Authentication : For storing user details like name, email, username, password we are using mongodb database. Sign Up routes have been set for user registration(post route), user details to be shown ( all user details and user details by id) and other CRUD routes such as updating ( patch to update user details) and delete. As it is a simple mobile application bcrypt js has been used for authentication rather than passport or oauth.
  2. Storing Labels: As soon as the model predicts the image either for landmark or object detection these details are stored in the database to study user images and its classification (CRUD routes have been set). This is important as based on more user activity we will improve our ML model as we will get more training and testing data.We store user details along with the picture( URL ) and label in the collection of the database. This is important as it will help us study user images.
  3. User feedback: After model prediction , user will fill the feedback form so that we can improve the ML model and it is stored in the database. ( CRUD routes have been set here again). This basically stores the user feedback on the model results, so that the ML model can be enhanced and provide better results.

Entire Project Out Soon!