Please use this identifier to cite or link to this item:
|Title:||Who is here: Location aware face recognition||Authors:||Wang, Z.
|Issue Date:||2012||Citation:||Wang, Z.,Yan, J.,Pang, C.,Chu, D.,Aghajan, H. (2012). Who is here: Location aware face recognition. Proceedings of the 3rd International Workshop on Sensing Applications on Mobile Phones, PhoneSense 2012 - In Conjunction with ACM SenSys 2012. ScholarBank@NUS Repository. https://doi.org/10.1145/2389148.2389156||Abstract:||Face recognition has many challenges. For instance, the illumination, various facial expression and different viewpoints add difficulties to identify the same person from a bunch of images. Searching over a huge set of images will only amplify such difficulties. We introduce the location aware face recognition framework for mobile-taken photos to alleviate the hardness. With the help of location sensor on the mobile devices, we collect images with location information. We propose an algorithm to reduce the search space of face recognition and therefore achieve better accuracy. Photos are clustered by locations on the server. Each location is then associated with a face classifier. Every client can send a "Who is Here" type query to the server by uploading an image with the location. The algorithm on the server will search over the given location and identify the person on the image. Experiments are conducted on mobile devices. The results are quite promising that higher accuracy is achieved and the query can be answered in near real-time. Copyright © 2012 ACM.||Source Title:||Proceedings of the 3rd International Workshop on Sensing Applications on Mobile Phones, PhoneSense 2012 - In Conjunction with ACM SenSys 2012||URI:||http://scholarbank.nus.edu.sg/handle/10635/125111||ISBN:||9781450317788||DOI:||10.1145/2389148.2389156|
|Appears in Collections:||Staff Publications|
Show full item record
Files in This Item:
There are no files associated with this item.
checked on Dec 4, 2021
checked on Dec 2, 2021
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.