Deep Learning in Photography

Last month, the Deep Learning Summit was held here in Boston and I was invited to speak. There were around 40 speakers and some incredible talks were given including Miguel Campo of 20th Century Fox, Anoop Deoras of Netflix, and Daniel Smilkov of Google Brain - definitely check them out! 

I gave a talk entitled Deep Learning Applications in Computational Photography, which is linked above, and the abstract/slides are below. If you don't have time to read it, VentureBeat wrote an article which gives a nice digest of my talk.

Abstract: Within the last year, there have been many advancements in the field of computational photography through the use of deep convolutional neural networks. Recently, techniques such as super resolution, style transfer, and modality substitution have revolutionized the state-of-the-art in professional imaging and computer vision. In this talk, a selection of new and recently published deep learning algorithms and projects for image processing are presented. In addition, we will surveying select new architectures that are very promising for photographic operations and examine some neural approaches to image filtration using fully-convolutional networks.