TY - GEN
T1 - Indoor Navigation using Text Extraction
AU - Eden, Jake
AU - Kawchak, Thomas
AU - Narayanan, Vijaykrishnan
N1 - Funding Information:
ACKNOWLEDGMENTS This work was supported in part by NSF Expeditions in Computing Program: Visual Cortex on Silicon CCF 1317560 and the SRC JUMP Center for Brain Inspired Computing. The views and findings expressed in this work are those of the author and not associated with the funding agencies.
PY - 2018/12/31
Y1 - 2018/12/31
N2 - In this paper, we explore the possibility of harnessing text extraction to perform localization for a visually impaired grocery shopper to help the shopper navigate through a grocery store. The extraction of environmental text is performed only after more traditional guidance techniques are employed to first bring the user to the text-rich aisles of a grocery store. This idea is built upon the need for both text extraction and localization in a visual assistance pipeline. For instance, typical visual assistance pipelines attempt to solve a number of problems: indoor localization and navigation, classification of an aisle's products, and potentially contextual information retrieval based on local environmental text (to, say, determine a product's price). However, prior art does not consider how text might also augment localization. Thus, this paper explores the viability of introducing text into such a seemingly disjoint problem space and ultimately concludes that environmental text extraction can enhance indoor localization by providing course-grained accuracy.
AB - In this paper, we explore the possibility of harnessing text extraction to perform localization for a visually impaired grocery shopper to help the shopper navigate through a grocery store. The extraction of environmental text is performed only after more traditional guidance techniques are employed to first bring the user to the text-rich aisles of a grocery store. This idea is built upon the need for both text extraction and localization in a visual assistance pipeline. For instance, typical visual assistance pipelines attempt to solve a number of problems: indoor localization and navigation, classification of an aisle's products, and potentially contextual information retrieval based on local environmental text (to, say, determine a product's price). However, prior art does not consider how text might also augment localization. Thus, this paper explores the viability of introducing text into such a seemingly disjoint problem space and ultimately concludes that environmental text extraction can enhance indoor localization by providing course-grained accuracy.
UR - http://www.scopus.com/inward/record.url?scp=85061403100&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85061403100&partnerID=8YFLogxK
U2 - 10.1109/SiPS.2018.8598409
DO - 10.1109/SiPS.2018.8598409
M3 - Conference contribution
AN - SCOPUS:85061403100
T3 - IEEE Workshop on Signal Processing Systems, SiPS: Design and Implementation
SP - 112
EP - 117
BT - Proceedings of the IEEE Workshop on Signal Processing Systems, SiPS 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2018 IEEE Workshop on Signal Processing Systems, SiPS 2018
Y2 - 21 October 2018 through 24 October 2018
ER -