Closed sanjeevghimire closed 5 years ago
@sanjeevghimire @stevemart Can you explain the motivation behind removing cloudant from this application?
@christiancompton TL;DR: much simpler setup, and less code (+121 −396) to maintain.
The point of the app is to showcase ARKit, CoreML, and Watson VisRec. We get the visual recognition data from photos to create the CoreML mode. The last step is to use ARKit to display some meaningful information, we initially thought this app would just displaying the confidence score from Watson VisRec, but decided to give it more information with sample "resume data" in a database. The feedback is that the information shown will differ between use cases and just makes a long setup that much more cumbersome (it requires a curl call for each entry).
@christiancompton In addition to what Steve said, there were numerous calls to Cloudant when the camera detects a face, to get the info back from cloudant which used up the lite quota in no time. This solution would address that as well.