Indoor Space Recognition using Deep Convolutional Neural Network: A Case Study at MIT Campus

Abstract : Global Position Systems and other navigation systems that collect spatial data through an array of sensors carried on by people and distributed in space have changed the way we navigate complex environments, such as cities. However, indoor navigation without reliable GPS signals relies on wall-mounted antennas, WiFi, or quantum sensors. Despite the gains of such technologies, underlying these navigation systems is the dismissal of the human wayfinding ability based on visual recognition of spatial features. In this paper, we propose a robust and parsimonious approach using Deep Convolutional Neural Network (DCNN) to recognize and interpret interior space. DCNN has achieved incredible success in object and scene recognition. In this study we design and train a DCNN to classify a pre-zoning indoor space, and from a single phone photo to recognize the learned space features, with no need of additional assistive technology. We collect more than 600,000 images inside MIT campus buildings to train our DCNN model, and achieved 97.9% accuracy in validation dataset and 81.7% accuracy in test dataset based on spatial-scale fixed model. Furthermore, the recognition accuracy and spatial resolution can be potentially improved through multiscale classification model. We identify the discriminative image regions through Class Activating Mapping (CAM) technique, to observe the model's behavior in how to recognize space and interpret it in an abstract way. By evaluating the results with misclassification matrix, we investigate the visual spatial feature of interior space by looking into its visual similarity and visual distinctiveness, giving insights into interior design and human indoor perception and wayfinding research. The contribution of this paper is threefold. First, we propose a robust and parsimonious approach for indoor navigation using DCNN. Second, we demonstrate that DCNN also has a potential capability in space feature learning and recognition, even under severe appearance changes. Third, we introduce a DCNN based approach to look into the visual similarity and visual distinctiveness of interior space.
Type de document :
Article dans une revue
PLoS ONE, Public Library of Science, 2016
Liste complète des métadonnées

Littérature citée [33 références]  Voir  Masquer  Télécharger
Contributeur : Dimitrios Milioris <>
Soumis le : vendredi 17 février 2017 - 14:48:37
Dernière modification le : lundi 8 janvier 2018 - 12:10:01
Document(s) archivé(s) le : jeudi 18 mai 2017 - 14:34:30


Fichiers produits par l'(les) auteur(s)


  • HAL Id : hal-01470497, version 1
  • ARXIV : 1610.02414



Fan Zhang​, Fabio Duarte​, Ruixian Ma, Dimitrios Milioris, Hui Lin​, et al.. Indoor Space Recognition using Deep Convolutional Neural Network: A Case Study at MIT Campus. PLoS ONE, Public Library of Science, 2016. 〈hal-01470497〉



Consultations de la notice


Téléchargements de fichiers