Development for google glass (preferable) or Vuzix (alternative).
The process is :
1) The user takes a picture and it is sent for processing
2) The processing returns the category of the picture (a string)
3) The category is displayed for the user to see
The development required is for steps 1 and 3. Step 2 is already developed in python (I can convert it to Tesnsorflow Lite or Android APK if needed, but from your part, no development would be required).
This is to be run on the device itself (if that is not possible, then it will use the cloud)
Draft diagram of the project is attached
User interface is simple. Just take the picture, no need to display the picture taken, then the classifier returns a string with the classification.
Not sure if the classifier will fit into the device, therefore need 2 quotes, one for 'in device' and the other 'using cloud services'.
Disregard the 'estimated budget' I filled up, I just put something because it is mandatory, I better would rather you to let me know your cost.