Image Aesthetic Distribution Prediction with Fully Convolutional Network

Huidi Fang, Chaoran Cui, Xiang Deng, Xiushan Nie, Muwei Jian, Yilong Yin

Abstract: Image aesthetics assessment emerges as a hot topic in recent years for its potential in numerous applications. In this paper, we propose to quantify the image aesthetics by a distribution over multiple quality levels. The distribution representation can effectively characterize the disagreement among users’ aesthetic perceptions regarding the same image. We realize an end-to-end framework of aesthetic distribution prediction with fully convolutional network, which accepts input images of arbitrary sizes. In this way, we circumvent the requirement of fixed-sized inputs from prevalent convolutional neural network, and thereby avoid the risk of impairing the intrinsic aesthetic appeal of images. Experiments on two benchmark datasets well verified the effectiveness of our approach in both scenarios of aesthetic distribution prediction and aesthetic label prediction.