I am currently teaching myself data science with a focus on bioinformatics. I have no formal training in anything like this. I am a self taught hardware and computer systems engineer.

I am working with a practice facial recognition challenge. My thoughts are that what I can learn from this can be use in developing approaches to image recognition with regard to pathology slide analyzation.

What I am trying to construct is this. There are almost "golden" ratios in the make up of a face.

Kind of like what's on this page - but not in the interest of defining beauty

Here is shows more of what I am thinking: http://www.goldennumber.net/face/

In order to recognize that is it s a face, I think I want to measure relationships between multiple points or factors nin the image.

Same as I might want to do with a pathology slide, looking for a particular pattern, or relationship of patterns on it.

I just don't know how to approach it, to break it down into something I can build back up in R. I am a novice at R as well.