See, caption, cluster: Large-scale image analysis using captioning and topic modeling

Kyeongpil Kang, Kyohoon Jin, Soojin Jang, Jaegul Choo, Youngbin Kim

Abstract: Highlights•We propose a novel topic modeling approach by combining text and visual contents via image captioning.•We propose a simple-but-robust scoring model to assess the quality of image captions generated from the image captioning model.•We summarize global information from the set of images, facilitating an intuitive understanding of the image dataset.