Autonomous learning for face recognition in the wild via ambient wireless cues
Facial recognition is a key enabling component for emerging Internet of Things (IoT) services such as smart homes or responsive offices. Through the use of deep neural networks, facial recognition has achieved excellent performance. However, this is only possibly when trained with hundreds of images...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Conference item |
| Published: |
Association for Computing Machinery
2019
|
| _version_ | 1826278063244050432 |
|---|---|
| author | Lu, C Kan, X Du, B Chen, C Wen, H Markham, A Trigoni, N Stankovic, J |
| author_facet | Lu, C Kan, X Du, B Chen, C Wen, H Markham, A Trigoni, N Stankovic, J |
| author_sort | Lu, C |
| collection | OXFORD |
| description | Facial recognition is a key enabling component for emerging Internet of Things (IoT) services such as smart homes or responsive offices. Through the use of deep neural networks, facial recognition has achieved excellent performance. However, this is only possibly when trained with hundreds of images of each user in different viewing and lighting conditions. Clearly, this level of effort in enrolment and labelling is impossible for wide-spread deployment and adoption. Inspired by the fact that most people carry smart wireless devices with them, e.g. smartphones, we propose to use this wireless identifier as a supervisory label. This allows us to curate a dataset of facial images that are unique to a certain domain e.g. a set of people in a particular office. This custom corpus can then be used to finetune existing pre-trained models e.g. FaceNet. However, due to the vagaries of wireless propagation in buildings, the supervisory labels are noisy and weak. We propose a novel technique, AutoTune, which learns and refines the association between a face and wireless identifier over time, by increasing the inter-cluster separation and minimizing the intra-cluster distance. Through extensive experiments with multiple users on two sites, we demonstrate the ability of AutoTune to design an environment-specific, continually evolving facial recognition system with entirely no user effort. |
| first_indexed | 2024-03-06T23:38:20Z |
| format | Conference item |
| id | oxford-uuid:6e6dfc4e-d3ee-4a27-b0ad-b74bf46b5ba2 |
| institution | University of Oxford |
| last_indexed | 2024-03-06T23:38:20Z |
| publishDate | 2019 |
| publisher | Association for Computing Machinery |
| record_format | dspace |
| spelling | oxford-uuid:6e6dfc4e-d3ee-4a27-b0ad-b74bf46b5ba22022-03-26T19:24:24ZAutonomous learning for face recognition in the wild via ambient wireless cuesConference itemhttp://purl.org/coar/resource_type/c_5794uuid:6e6dfc4e-d3ee-4a27-b0ad-b74bf46b5ba2Symplectic Elements at OxfordAssociation for Computing Machinery2019Lu, CKan, XDu, BChen, CWen, HMarkham, ATrigoni, NStankovic, JFacial recognition is a key enabling component for emerging Internet of Things (IoT) services such as smart homes or responsive offices. Through the use of deep neural networks, facial recognition has achieved excellent performance. However, this is only possibly when trained with hundreds of images of each user in different viewing and lighting conditions. Clearly, this level of effort in enrolment and labelling is impossible for wide-spread deployment and adoption. Inspired by the fact that most people carry smart wireless devices with them, e.g. smartphones, we propose to use this wireless identifier as a supervisory label. This allows us to curate a dataset of facial images that are unique to a certain domain e.g. a set of people in a particular office. This custom corpus can then be used to finetune existing pre-trained models e.g. FaceNet. However, due to the vagaries of wireless propagation in buildings, the supervisory labels are noisy and weak. We propose a novel technique, AutoTune, which learns and refines the association between a face and wireless identifier over time, by increasing the inter-cluster separation and minimizing the intra-cluster distance. Through extensive experiments with multiple users on two sites, we demonstrate the ability of AutoTune to design an environment-specific, continually evolving facial recognition system with entirely no user effort. |
| spellingShingle | Lu, C Kan, X Du, B Chen, C Wen, H Markham, A Trigoni, N Stankovic, J Autonomous learning for face recognition in the wild via ambient wireless cues |
| title | Autonomous learning for face recognition in the wild via ambient wireless cues |
| title_full | Autonomous learning for face recognition in the wild via ambient wireless cues |
| title_fullStr | Autonomous learning for face recognition in the wild via ambient wireless cues |
| title_full_unstemmed | Autonomous learning for face recognition in the wild via ambient wireless cues |
| title_short | Autonomous learning for face recognition in the wild via ambient wireless cues |
| title_sort | autonomous learning for face recognition in the wild via ambient wireless cues |
| work_keys_str_mv | AT luc autonomouslearningforfacerecognitioninthewildviaambientwirelesscues AT kanx autonomouslearningforfacerecognitioninthewildviaambientwirelesscues AT dub autonomouslearningforfacerecognitioninthewildviaambientwirelesscues AT chenc autonomouslearningforfacerecognitioninthewildviaambientwirelesscues AT wenh autonomouslearningforfacerecognitioninthewildviaambientwirelesscues AT markhama autonomouslearningforfacerecognitioninthewildviaambientwirelesscues AT trigonin autonomouslearningforfacerecognitioninthewildviaambientwirelesscues AT stankovicj autonomouslearningforfacerecognitioninthewildviaambientwirelesscues |