作者: Pavel Král , Antonín Vrba , Ladislav Lenc
DOI: 10.1007/978-3-030-20915-5_3
关键词: Local binary patterns 、 Benchmark (computing) 、 Artificial intelligence 、 Facial recognition system 、 Pattern recognition 、 Image resolution 、 Image noise 、 Feature vector 、 Computer science 、 Face (geometry) 、 Pixel
摘要: This paper presents a novel automatic face recognition approach based on local binary patterns. descriptor considers neighbourhood of pixel to compute the feature vector values. method is not very robust handle image noise, variances and different illumination conditions. We address these issues by proposing which more pixels neighbourhoods The proposed evaluated two benchmark corpora, namely UFI FERET datasets. experimentally show that our outperforms state-of-the-art methods efficient particularly in real conditions where above mentioned are obvious. further handles well one training sample issue also resolution.