Investigate available camera capabilities we can leverage

[azasypkin]

@sfoster @punamdahiya feel free to add ideas and capabilities you think would be useful to have.

I see if we can rely on something that tells us that we have an object in focus, not blurred etc. Otherwise maybe I'll check if we can use OpenCV for that (eg. detect contours, extract keypoints and see if we have enough of them).

[azasypkin]

https://developer.apple.com/library/content/documentation/DeviceInformation/Reference/iOSDeviceCompatibility/Cameras/Cameras.html

Strangely (or obviously since we're talking about Apple) iPhone 5(s) cameras aren't mentioned.

[azasypkin]

Camera properties we can control directly from C++ (VideoCapture): https://github.com/opencv/opencv/blob/master/modules/videoio/include/opencv2/videoio.hpp#L486-L491

[azasypkin]

Okay here is the list of all currently available frame metadata that iOS can give us. Basically face, none and a bunch of machine readable codes (that we can use later on):

• AztecCode
• Code128Code
• Code39Code
• Code39Mod43Code
• Code93Code
• DataMatrixCode
• EAN13Code
• EAN8Code
• Face
• Interleaved2of5Code
• ITF14Code
• None
• PDF417Code
• QRCode
• UPCECode

[sfoster]

That list means nothing to me. Can you summarize the significance of these metadata properties @azasypkin

[azasypkin]

That list means nothing to me.

Well the summary is pretty short - there is nothing useful for us at this stage in iOS framework, we'll have to rely on something based on OpenCV.

AVFoundationFramework can only give us additional metadata about faces or machine readable codes it sees in the photo (location, bounding rect, roll/yaw angle and type for the codes). We may need it later though, but it's too early to think about it.