2020/11/26 by Samadhi Wickrama Arachchilage, Arachchilage, S. W., Ebroul Izquierdo +1
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Face and Expression Recognition #Face recognition and analysis #I.4.0 #I.5.0 #Machine Learning (cs.LG)
paper · pdf · doi:10.48550/arxiv.2011.13361
openalex publication_date 2020/11/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/30
Face recognition in unconstrained environments is challenging due to variations in illumination, quality of sensing, motion blur and etc. An individual's face appearance can vary drastically under different conditions creating a gap between train (source) and varying test (target) data. The domain gap could cause decreased performance levels in direct knowledge transfer from source to target. Despite fine-tuning with domain specific data could be an effective solution, collecting and annotating data for all domains is extremely expensive. To this end, we propose a self-supervised domain learning (SSDL) scheme that trains on triplets mined from unlabelled data. A key factor in effective discriminative learning, is selecting informative triplets. Building on most confident predictions, we follow an "easy-to-hard" scheme of alternate triplet mining and self-learning. Comprehensive experiments on four different benchmarks show that SSDL generalizes well on different domains.