Skip to content

Latest commit

 

History

History
23 lines (18 loc) · 1.01 KB

notes.md

File metadata and controls

23 lines (18 loc) · 1.01 KB

Backend

Django

  • Why?
    • Familiarity with django
      • Two weeks is not enough time to learn a new framework
  • Speed
    • Like any good framework, most things are done for you! For the demo, we didn't have to worry about
      • hosting
      • database management
      • user authentication
    • Can develop extremely fast
    • Very well documented

Machine learning

Demo Limitations. Trade off between accuracy and performance (can't use models that are too computationally expensive for our mobile application). Problem: Not enough time to gather food dataset, and to train a model. Solution: pretrained Pretrained on the imagenet classification challenge (1000 classes). Input is an image, output is a "confidence" value for each of the 1000 classes. Problem: Of these 1000 classes, only around 50 are food, don't want to be able to classify non-food images. Solution: Only look at the outputs of food classes to determine the most likely food, and normalize the confidence of these classes outputs.