Abstract
Narrative collage is an interesting image editing art to summarize the main theme or storyline behind an image collection. We present a novel method to generate narrative images with plausible semantic scene structures. To achieve this goal, we introduce a layer graph and a scene graph to represent relative depth order and semantic relationship between image objects, respectively. We firstly cluster the input image collection to select representative images, and then extract a group of semantic salient objects from each representative image. Both Layer graphs and scene graphs are constructed and combined according to our specific rules for reorganizing the extracted objects in every image. We design an energy model to appropriately locate every object on the final canvas. Experiment results show that our method can produce competitive narrative collage result and works well on a wide range of image collections.
Funding Information
  • the national key Research and Development Program of China (2017YFB1002600)
  • NSFC (61472288, 61672390)
  • Foundation of Key Research Institute of Humanities and Social Science at Universities (16JJD870002)
  • Chinese Ministry of Education
  • Wuhan Science and Technology Plan Project (2017010201010109)

This publication has 34 references indexed in Scilit: