“A Deep Architecture for Unified Esthetic Prediction”, 2017-08-16 (; backlinks; similar):
Image esthetics has become an important criterion for visual content curation on social media sites and media content repositories. Previous work on esthetic prediction models in the computer vision community has focused on esthetic score prediction or binary image labeling. However, raw esthetic annotations are in the form of score histograms and provide richer and more precise information than binary labels or mean scores. Consequently, in this work we focus on the rarely-studied problem of predicting esthetic score distributions and propose a novel architecture and training procedure for our model.
Our model achieves state-of-the-art results on the standard AVA large-scale benchmark dataset for 3 tasks: (1) esthetic quality classification; (2) esthetic score regression; and (3) esthetic score distribution prediction, all while using one model trained only for the distribution prediction
task. We also introduce a method to modify an image such that its predicted esthetics changes, and use this modification to gain insight into our model.