Video-to-Video Face Recognition for Low-Quality Surveillance DataThe availability of video data is an opportunity and a challenge for law enforcement agencies. Face recognition methods can play a key role in the automated search for persons in the data. This work targets efficient representations of low-quality face sequences to enable fast and accurate face search. Novel concepts for multi-scale analysis, dataset augmentation, CNN loss function, and sequence description lead to improvements over state-of-the-art methods on surveillance video footage. |
Other editions - View all
Video-to-Video Face Recognition for Low-Quality Surveillance Data Christian Herrmann No preview available - 2020 |
Video-to-Video Face Recognition for Low-Quality Surveillance Data Christian Herrmann No preview available - 2020 |
Common terms and phrases
acc std applications arXiv arXiv preprint augmentation strategy average precision bag-of-words baseline camera center-based ChokePoint classical comparison Computer Vision Conference on Computer convolutional layer data quality denotes dense SIFT domain Euclidean distance Face Database face datasets face descriptor face image descriptors face recognition face retrieval face samples face sequence descriptor face track face verification FaceScrub feature filter footage fully connected layers head pose HERRMANN high-quality histogram histogram fusion hypersphere identity IEEE includes index database indices input International Conference inverted index inverted index strategy IOSB-SURV ISBN learning loss function low-resolution LqfNet LR-LBP matching max-margin methods minset motion blur multi-scale Neural Networks neurons overfitting parameters Pattern Recognition performance pixels pooling layers proposed query radius region residual resolution robust Sequence representation supervised strategy target space tion track descriptor unsupervised validation vector VGG-Face video data video face Vision and Pattern visual words