First Workshop on

Large-scale Scene Reconstruction and Visual Mapping using Crowdsourcing Images.


The goal of the workshop is to present and discuss state of the art techniques, in the field of multidimensional scene reconstruction based on large datasets. It provides an opportunity to debate on current main challenges in this field, as well as to establish cooperation with other researchers interested in the domain.


  • sparse and dense scene reconstruction techniques(different approaches for SfM, alternatives to CMVS/PMVS)
  • scalability improvements
  • popular open-source tools for 3D reconstruction (Bundler, VisualSFM, PMVS/CMVS)
  • improving visual quality of the reconstruction
  • temporal changes detection
  • applications related to processing large image datasets
  • current limitations

This workshop calls for paper submissions concerning following above subjects and fields related.


The emergence and increasing popularity of numerous public image sharing websites like instagram or flickr and also services like google street-view, gave the researchers the possibility to acquire large image data sets in easy way. The natural consequence of that was development of techniques that could automatically process the huge amount of data and leverage numerous possibilities that crowdsourcing gives us like automatic modeling of famous monuments or building 3D city maps just to name a few . The aim of this workshop is to present most recent techniques used for operating on crowdsourcing image data as well as different applications that take advantage of this specific data source. As this is still a young domain there is a lot of room for improvement in terms of quality and scalability of reconstruction, as well as in the ways the output of this techniques is utlized(e.g. city scale change detection, image based search in 3D models etc.). This workshop is design for those who already have an experience with this domain as well as for those who want to learn about most important ideas and state of the art related to the topic. It will allow to collectively name the most important challenges in the domain as well as set new directions for development.