Iris is an image analysis program that given a manifest file, and a set of images will return the relative positions of each capturing camera from one another. For example the follwowin two images were taken by two phones beside each other. This one on the left: left
image and this one on the right: right

So how do i find which picture is which? Well they are very close to one another so I then used this to my advantage to use opencv to find the matching landmarks between the two images. It then returns the following image. comparisons

Now that I have found the matching points in the two images I can find there distances from the left side of the image. This then tells me there relative positions. This is because the left camera will see the same landmarks further to the left and the right camera will see the landmarks further to the right. This seems anti intuitive especially if you try moving your head left and right while staring at an object you will see that on the left the object will appear furhter to the right. The reason the code is the other way around is that images are mirrored by the front camera phones.

So there we have it, at the code will return a "matrix" of scores that are then used by the backend code to crop and position an image or video: {'scrs': [{'ip2': u'', 'ip1': u'', 'score': 0.0}, {'ip2': u'', 'ip1': u'', 'score': 65.76188568516939}, {'ip2': u'', 'ip1': u'', 'score': -17.09808192707039}, {'ip2': u'', 'ip1': u'', 'score': 241.95999999999995}]}