LauzHack 2018
IBM api:
pip install --upgrade "watson-developer-cloud>=2.4.1"
Idea/goals:
- Every 30 seconds: take 4K picture on client
- blur people on client
- send blurred and patched versions from client to server (= increased security/anonymity)
- Server does API call to IBM Watson Vision to classify as food/non-food
- Make snips.ai board talk with you, that you should go and check out food in #building. You can then ask what food is present.
- Store results in database: (timestamp, building, {11: "food", 12: "non-food", ...}). keys are array indexes ij of patches
- For those patches that come back with "food" or "fruit" or ..., send them to the IBM Watson Vision "food" API to determine the type of food. Store this too.
- Count people