After multiple attempts, with various light sources and times of day, the best model came out when we used morning indoor lighting that was diffused by curtains. A problem that I ran into on various attempts too was when the subjects hair was down because it created a solid block that did not have much variation in contrasts. I also tried an inanimate abject to see if that would work better. From this attempt I found that it worked best when the most pictures possible were taken but the app on the phone limits that, which resulted in the from being the most accurate part of the model.
