Video2Book: semi-automatic tool to create mediaproducts from videos on smart phones

  • Authors:
  • Mohamad Rabbath;Volker Gollücke;Susanne Boll

  • Affiliations:
  • OFFIS (Oldenburg Institute of Technology), Oldenburg, Germany;University of Oldenburg, Oldenburg, Germany;University of Oldenburg, Oldenburg, Germany

  • Venue:
  • Proceedings of the 3rd ACM conference on International conference on multimedia retrieval
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Users nowadays increasingly use their smart phones to take photos and videos. Many of the users would like to create multimedia products such as physical photo books out of their photos or videos on the phone, however the screen of the phone is small for complicated editing and publishing process. In this demo paper we introduce an application that allows the user to automatically create a photo product out of his videos on the smart phone. The user can select one or several videos. The scenes of each video are detected using visual features, and for each scene a distinct good quality frame is selected. The videos are usually not very long and on average a video is 2 mins length in our dataset of users. We enrich the extracted representing frames with related photos from both the phone and from social networks like Facebook. We measure the similarity based on visual, metadata and social features. We apply a face recognition phase to enhance the accuracy of these features. At the end we compile the selected media content in an appealing photo product, that the user can edit later from any device. This process allows the user to make use of the phone videos to create appealing media products with very little effort that suits the small screen of the mobile phone.