Constrained Nonnegative Matrix Factorization for Image Representation

By Haifeng Liu, Zhaohui Wu, Xuelong Li, Deng Cai, Thomas S. Huang. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012

505 citations recorded by the source corpus.

Nonnegative matrix factorization (NMF) is a popular technique for finding parts-based, linear representations of nonnegative data. It has been successfully applied in a wide range of applications such as pattern recognition, information retrieval, and computer vision. However, NMF is essentially an unsupervised method and cannot make use of label information. In this paper, we propose a novel semi-supervised matrix decomposition method, called Constrained Nonnegative Matrix Factorization (CNMF), which incorporates the label information as additional constraints. Specifically, we show how explicitly combining label information improves the discriminating power of the resulting matrix decomposition. We explore the proposed CNMF method with two cost function formulations and provide the corresponding update solutions for the optimization problems. Empirical experiments demonstrate the effectiveness of our novel algorithm in comparison to the state-of-the-art approaches through a set of evaluations based on real-world applications.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/2c024a6b7d5ab679f88202e413d5399dc8414257

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Dacheng Tao
2. Hanwang Zhang
3. Luc Van Gool
4. Shuicheng Yan
5. Tat-seng Chua
6. Yi Yang
7. Tao Mei
8. Xiaogang Wang

Citations in the corpus

• Robust (Semi) Nonnegative Graph Embedding (2014, background): “[17] proposed a semi-supervised NMF algorithm that incorporates label information in the coefficient matrix.”
• A Deep Matrix Factorization Method for Learning Attribute Representations (2015, methodology): “CNMF [24], DNMF [25], and our WSF models that have attribute labels only for the attribute we were classifying for, and our WSF-MA and Deep WSF that learned data representations based on all attribute information available.”
• Deep Collaborative Embedding for Social Image Understanding (2019, background): “In [52], CNMF is designed by forcing images with the same label to have the same latent representations.”