Using Deep Learning to Create Professional-Level Photographs

Discussion in '[H]ard|OCP Front Page News' started by Megalith, Jul 16, 2017.

  1. Megalith

    Megalith 24-bit/48kHz Staff Member

    Messages:
    10,488
    Joined:
    Aug 20, 2006
    Google has taught an AI to edit photos like a professional, and the results are pretty interesting: the steps included taking roaming landscape panoramas from Google Street View, searching for the best composition, then carrying out various postprocessing operations to create an aesthetically pleasing image. The full gallery may be found here.

    Our approach relies only on a collection of professional quality photos, without before/after image pairs, or any additional labels. It breaks down aesthetics into multiple aspects automatically, each of which is learned individually with negative examples generated by a coupled image operation. By keeping these image operations semi-”orthogonal”, we can enhance a photo on its composition, saturation/HDR level and dramatic lighting with fast and separable optimizations.