How to store and query biometric data for an authentication system?

I am trying to design, and hopefully implement, an authentication system which centers around the use of biometric images. I plan to use different machine learning and deep learning techniques to help with this through Tensorflow. Specifically, the project will focus on the use of image processing for authentication.

I know that in a traditional authentication system, for example using a username and password combination, both data are stored in a database with a specific data type (usually strings). When verifying the user, we could just query the username and match the hashed password. This is pretty simple because we are storing structured data. However, with biometrics, data are usually unstructured usually in the form of images. It is difficult to efficiently store unstructured data and it is even more difficult to match a stored unstructured data with a new one for testing. It is just too computationally extensive to test the set of stored data, one by one, to the new input data.

One idea I had is to generate a function taking these images as input and output a number which would act as an identifier for that specific image. It is like a function which places images in a vector space and the images can be matched when they have the similar (or close to similar) value in the vector space. Is this appropriate approach with my use case? However, I am still having problem how and where will I store my data.

I am still a beginner in terms of data science so it will gladly help if you could recommend me further readings or similar projects.