In recent years, nonnegative matrix factorization (NMF) methods of a reduced image data representationattracted the attention of computer vision community. These methods are considered as a convenient part-basedrepresentation of image data for recognition tasks with occluded objects. A novel modification in NMFrecognition tasks is proposed which utilizes the matrix sparseness control introduced by Hoyer. We haveanalyzed the influence of sparseness on recognition rates (RRs) for various dimensions of subspaces generatedfor two image databases, ORL face database, and USPS handwritten digit database. We have studied thebehavior of four types of distances between a projected unknown image object and feature vectors in NMF subspacesgenerated for training data. One of these metrics also is a novelty we proposed. In the recognitionphase, partial occlusions in the test images have been modeled by putting two randomly large, randomlypositioned black rectangles into each test image.
展开▼
