A novel constraint non-negative matrix factorization criterion based incremental learning in face recognition

This paper addresses incremental learning and time-consuming problems in non-negative matrix factorization (NMF) of face recognition. When the training samples or classes are incremental, almost all existing NMF based methods must implement repetitive learning. Also, they are usually very time-consuming. To overcome these limitations, we proposed a novel constraint block NMF (CBNMF) method, which is based on a new constraint NMF criterion and our previous block technique in NMF. CMU PIE face database is selected for evaluation. Comparing with Block NMF (BNMF), NMF and PCA methods, experimental results show that our proposed CBNMF approach gives the best performance.