Abstract
Efficient encoding of facial descriptors remains to be a major topic for face recognition. Among various methods, Fisher vector (FV) representations have shown satisfying performance on most benchmark datasets. However, its representation is huge. In this paper, we present a novel approach to make Fisher vector compact and improves its performance. We utilize handcrafted low-level descriptors as FV do. However, we retain only 1st order statistics of FV, introduce Gaussian block to sparsify FV, alter its formulation, and normalize properly. We evaluate our method on LFW and FERET dataset, and result shows our method effectively compresses Fisher vector and achieves satisfying result at the same time.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
