Using A New Approach in Deep Dictionary Learning to Handwriting Number Classification

Dictionary learning and sparse representation is a successful mathematical model for data representation that achieves state-of-the-art performance in various fields such as pattern recognition, machine learning, computer vision. The study aims to improve the classification performance of state-of-the-art methods by using a multi- layer framework. This paper presents the new idea of "multi- layered K-singular value decomposition (MLK-SVD)" dictionary learning as a multi-layer method of classification and is evaluated on MNIST dataset. The new idea is a multi- layer dictionary learning method for classification tasks. In order to learn better features, the frame uses a tag consistent in the K-SVD algorithm to learn a discriminative dictionary for sparse coding. We also include information on labels in addition to class labels for training data.