NUST Institutional Repository

Pose Calibrated Feature Aggregation for Robust Video Face Recognition in Unconstrained Environments

Show simple item record

dc.contributor.author Hasani, M Ibrahim Ali
dc.date.accessioned 2022-04-28T04:58:41Z
dc.date.available 2022-04-28T04:58:41Z
dc.date.issued 2022
dc.identifier.uri http://10.250.8.41:8080/xmlui/handle/123456789/29308
dc.description.abstract This paper presents PCFAN, a network architecture for set/video face recognition. Using stacked attention blocks and a multi-model architecture, it automatically assigns adaptive weights to every instance in the set, based on both the recognition embeddings and the associated face metadata. It uses these weights to produce a single, compact feature vector for the set. The model automatically learns to advocate for features from images with more favorable qualities and poses, which inherently hold more information. Our block can be inserted on top of any standard recognition model for set prediction and improved performance, particularly in unconstrained scenarios where subject pose and image quality vary considerably between frames. We test our approach on three challenging video face-recognition datasets, IJB-A, IJB-B, and YTF, and report state-of-the-art results. Moreover, a comparison with top aggregation methods as our baselines demonstrates that PCFAN is the superior approach. en_US
dc.description.sponsorship Dr. Omar Arif en_US
dc.language.iso en en_US
dc.publisher SEECS, National University of Sciences & Technology Islamabad en_US
dc.subject set recognition, unconstrained video face recognition, feature aggrega tion, adaptive weighing, m en_US
dc.subject demonstrates that PCFAN is the superior approach. Keywords: set recognition, unconstrained video face recognition, feature aggrega tion, adaptive weighing, multi-model networ en_US
dc.title Pose Calibrated Feature Aggregation for Robust Video Face Recognition in Unconstrained Environments en_US
dc.type Thesis en_US


Files in this item

This item appears in the following Collection(s)

  • MS [375]

Show simple item record

Search DSpace


Advanced Search

Browse

My Account