Publications / 2018 Proceedings of the 35th ISARC, Berlin, Germany

Image-based Indoor Localization Using BIM and Features of CNN

Inhae Ha, Hongjo Kim, Somin Park and Hyoungkwan Kim
Pages 773-776 (2018 Proceedings of the 35th ISARC, Berlin, Germany, ISBN 978-3-00-060855-1, ISSN 2413-5844)
Abstract:

This study suggests an indoor localization method to estimate the location of a user of a mobile device with imaging capability. The proposed method uses a matching approach between an actual photograph and a rendered BIM (building information modeling) image. A pre-trained VGG 16 network is used for feature extraction. Experimental results show that the best image matching performance can be obtained when using features from pooling layer 4 of VGG16. The proposed method allows for indoor localization only by image matching without additional sensing information.

Keywords: Indoor Localization, Cross-Domain Image Matching, Convolutional Neural Network, Feature Extraction,