I have, as I am sure many others do, questions about how IQE works to determine its applicability for me and the products/services I would like to integrate it with. A few quick examples

- is it color sensitive? if i create an object with identical images that are different colors (say a blue ford focus and green ford focus) will it be able to match a yellow ford focus with that object?

- what are some tolerance guidelines for missing information or put another way is the hit/miss ratio for an image match available?

- if there are multiple objects in a photo and only one is returned how do we retrain it (i.e. associate it with multiple objects manually)

- how does your crowdsourcing platform work? is it a mechanical turk job? who has access to my private images? what if i am not satisfied with that interpretation?

- how do i associate new information with existing objects in a private db? for example my first attempt gave a label but i know want to add metadata (POST /object/update/<id> adds metadata, POST /object/<id> overwrites all metadata data with POST data.)

It would be nice if when pictures were added to the private database, they could be assigned an optional "category ID", which can be used in future queries so that only pictures that have the matching category ID are considered in the query.

Strategic use of this feature would speed up queries and reduce server load, and it would enable developers to have multiple isolated sets of images that should not be mixed (e.g. lets say you made a mobile app that multiple people will use, it might be useful for each person to have a private collection of images to search)