Using Caffe for Particle Physics

[Neil McFadden]

Hi,

I have been reading about machine vision while I have been working on my physics PhD and I think I can use Caffe to optimize my data analysis.

My biggest concern is how I could turn a physics data set into something Caffe can work with. The physics data set has individual events (like pixels) 

that have an arbitrary number of attributes, like momentum, energy and position (like the three channels of color). Similar to an image, the input layer to my net would be something 

like 1xnx1xN (versus 1x3xNxN), where n is the number of attributes for each event and N is the number of events in the data set. I imagine that getting my data set

into an image like format might be a nightmare because I would have to convert it from it's native format to something that is recognizable for Caffe.

Based on what I have said, does it seem like Caffe can be adapted to this problem and if not, does anyone know if there is a better tool for me to try?

Thanks,

Neil