What makes an Image Iconic? A Fine-Grained Case Study
This work addresses the problem of selecting representative images for visual concepts, which is incremental as it builds on existing iconicity research but applies it to a fine-grained domain.
The study investigated whether iconicity of images is subjective, predictable, and what factors contribute to it, using a fine-grained bird dataset, and found that iconicity is consistent across individuals and can be predicted by combining intuitive properties and a trained predictor to approach human performance.
A natural approach to teaching a visual concept, e.g. a bird species, is to show relevant images. However, not all relevant images represent a concept equally well. In other words, they are not necessarily iconic. This observation raises three questions. Is iconicity a subjective property? If not, can we predict iconicity? And what exactly makes an image iconic? We provide answers to these questions through an extensive experimental study on a challenging fine-grained dataset of birds. We first show that iconicity ratings are consistent across individuals, even when they are not domain experts, thus demonstrating that iconicity is not purely subjective. We then consider an exhaustive list of properties that are intuitively related to iconicity and measure their correlation with these iconicity ratings. We combine them to predict iconicity of new unseen images. We also propose a direct iconicity predictor that is discriminatively trained with iconicity ratings. By combining both systems, we get an iconicity prediction that approaches human performance.