# Intrinsic Image Popularity Assessment

## TL;DR

This paper introduces a large-scale database and deep learning models for intrinsic image popularity assessment, demonstrating superior performance and human-level accuracy in predicting social media image virality.

## Contribution

It presents the first large-scale database for intrinsic image popularity assessment and develops deep neural network models optimized for ranking consistency.

## Key findings

- Model outperforms existing methods on social media platforms.
- Achieves human-level performance on Instagram.
- Demonstrates good generalizability across different datasets.

## Abstract

The goal of research in automatic image popularity assessment (IPA) is to develop computational models that can accurately predict the potential of a social image to go viral on the Internet. Here, we aim to single out the contribution of visual content to image popularity, i.e., intrinsic image popularity. Specifically, we first describe a probabilistic method to generate massive popularity-discriminable image pairs, based on which the first large-scale image database for intrinsic IPA (I$^2$PA) is established. We then develop computational models for I$^2$PA based on deep neural networks, optimizing for ranking consistency with millions of popularity-discriminable image pairs. Experiments on Instagram and other social platforms demonstrate that the optimized model performs favorably against existing methods, exhibits reasonable generalizability on different databases, and even surpasses human-level performance on Instagram. In addition, we conduct a psychophysical experiment to analyze various aspects of human behavior in I$^2$PA.