I am making an app in flutter. I would like to implement a feature where user can upload a image from database and it goes through a series of previously uploaded images on the firebase storage and returns if there's a match for the people in the database. How would I go about implementing this, simple and fast. A way to just compare two images passed on to a method will also be fine.
3 Answers
My advice is: take a look on face_recognotion lib (based on dlib) just to get better understanding how to compare faces, what face vectorization is and how to compare that vectors (or so called "face encodings" or "face embeddings")). That understanding may push you to an idea that you need to store not only face image files but face embedding too to be able to compare them. Then you could start to think either your Firebase has in-built functionality to index and compare that face embeddings (vectors).. or you should find a db server that has (Elasticsearch, for example).. Then read about Google Mediapipe - probably the fastest framework for object recognition for mobile apps.
I've experience using Face recognition API from Luxand in my App. These guys (Luxand) are high-tech. It’s cool, their customers include Universal Pictures, Boeing, Samsung, etc. You can try it for free to understand the benefits of the Face recognition API for your App. More info here https://luxand.cloud/?utm_source=stackoverflow&utm_medium=socialstackoverflow
Related